The Sex Lives of Female Olive Baboons (Papio Anubis)

Home , Concealed ovulation, Olive baboon, Sexual swelling, Yellow baboon

Competition, coercion, and choice: The sex lives of female olive baboons (Papio anubis)

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy

in the Graduate School of The Ohio State University

Jessica Terese Walz

Graduate Program in Anthropology

The Ohio State University

2016

Dissertation Committee:

Dawn M. Kitchen, Chair

Douglas E. Crews

W. Scott McGraw

Copyrighted by

Jessica Walz

2016

Abstract

Since Darwin first described his theory of sexual selection, evolutionary biologists have used this framework to understand the potential for morphological, physiological, and behavioral traits to evolve within each sex. Recently, researchers have revealed important nuances in effects of sexual coercion, intersexual conflict, and sex role reversals. Among our closest relatives living in complex societies in which individuals interact outside of just the context of mating, the sexual and social lives of individuals are tightly intertwined. An important challenge to biological anthropologists is demonstrating whether female opportunities for mate choice are overridden by male- male competitive and male-female coercive strategies that dominate multi-male, multifemale societies.

In this dissertation, I explore interactions between these various mechanisms of competition, coercion, and choice acting on the lives of female olive baboons to determine how they may influence expression of female behavioral and vocal signals, copulatory success with specific males, and the role of female competition in influencing mating patterns. I found females solicit specific males around the time of ovulation.

Although what makes some males more preferred is less clear, there is evidence females choose males who might be better future protectors – males who will have long group tenures and are currently ascending the hierarchy. Preference translates into higher copulation rates and success at consort takeovers, there is little support that this is simply

ii based on male aggression toward females. Outside the fertile window female copulations were more likely related to male aggression and male dominance rank. Additionally, I found evidence that copulation calls of female olive baboons indicate ovulation and may function to encourage specific males to guard or continue mating with females. However, some of the temporal features of calls indicate a function for paternity confusion. Finally, cycling females were targeted for more aggression than they gave, suggesting aggression among females may limit reproductive competition. However, among ovulating females in consort with preferred males, most aggression was directed toward lactating females.

These are females who present the biggest threat for cycling females to develop a bond with a future male “friend.” In baboons, establishing nonsexual friendships is valuable in terms of mother and infant protection.

Overall, I showed that the social and sexual lives of female olive baboons differ as they approach periods in their ovulatory cycles when they are more likely to conceive, and provide support for hypotheses suggesting female strategies, like darting, copulation calls, and proceptive behaviors toward males with certain qualities appear to serve as a graded signal that allows females to both encourage mating from preferred partners when near conception, but also encourage competition or paternity confusion when not fertile.

Female strategies seem most focused on ensuring conception, but also may be important for establishing bonds with males for future protection. I also highlight important similarities and differences in male and female reproductive strategies across three savanna baboon species and consider how a synthesis of these unique strategies can help resolve questions about evolved sexual conflicts in the context of complex societies.

iii

For my family

Acknowledgements

The work that went into this dissertation could not have been completed without the encouragement, help, and support of many people. First, I would like to thank my advisor, Dawn Kitchen. You have been such an amazing mentor, supporter, and friend.

Your tireless energy and long phone conversations with words of encouragement and advice allowed me to push forward. I could not have done this without you. I also want to thank Doug Crews for your guidance through this process. You have always encouraged me and supported me in all my endeavors. Thanks also to Scott McGraw for pushing me to think critically and broadly about my perspective and approach and encouraging me to

“dock the ship.” Each of your teaching and mentoring styles helped me grow, not only as a researcher, but also as a college educator. Thank you also to Clark Larsen and The Ohio

State University Department of Anthropology for always supporting my research and teaching endeavors.

Thank you to my many financial contributors and institutions, especially The

Leakey Foundation, International Primatological Society, American Society of

Primatology, Animal Behavior Society, American Society of Mammalogists, both the national and the Ohio State University chapters of Sigma Xi, Ohio State’s Graduate

School, Office of International Affairs, and the Department of Anthropology.

To all of the people at Gombe Stream National Park who befriended me and assisted me in my work: asante sana! I am especially thankful to Anton Collins who I am

v proud and humbled to call a friend and mentor. Your limitless knowledge of the park, the baboons, and their habitat helped me complete my project in a comprehensive and methodical way. Thank you to Deus Mjungu for your guidance and assistance in my transition to life at Gombe. Thank you also to the team of baboon researchers, especially

Marini Bwenda. The many hours you spent helping me identify baboons and learn their ranges was integral to the success of my project. I could never have started this endeavor without our intense “baboon bootcamp.” Special thanks also to Andrea Bailey for being an amazing friend, supporter, and collaborator. Without your help I quite literally would have been lost in the woods. I am also forever indebted to Hashim Issa and Hamimu

Mbwama. Your ability to find and track baboons (and their “sampos”) constantly amazed me and I could not have completed this project without your help. Thanks also to Ashura

Issa for your friendship and masterful cooking skills. Coming home from a long day in the field to find a home-cooked meal waiting for me was often what helped me get through tough days in the field. Thanks to Iddi Lipende and Juma Baranyikwa for allowing me to share space in the Gombe lab, helping me find and purchase necessary supplies to complete hormone extractions, and weighing fecal samples. My time at

Gombe would not have been as enjoyable without the friendship of Kara Walker-

Schroepfer and Lisa O’Bryan. From sun-downers to commiserating about trials and tribulations of conducting fieldwork, your friendship helped shape my experience at

Gombe.

Thank you to Jacinta Beehner and Teera Parr and the rest of the Core Assay

Facility team. Housing and processing the hundreds of fecal samples I collected to

vi complete this project would not have been possible without your help. Thank you to

Jacinta for the many correspondences along the way and to Teera for spending time training me in important fecal hormone sample assay techniques.

All of my anthropology friends and friends outside of the field have also been incredibly supportive in this process. Special thanks go to Jill Murphey and Bryan

Johnson for always taking me in to your home and treating me like family. Liz Perrin

Beggrow (and Adam), Britney Kyle, and Laurie Reitsema: you are my anthropology

“sisters” and I could not have made it through coursework and life’s trials and tribulations without your support. Thank you also to Michelle Rodrigues for a formative experience in the Costa Rican rainforest that helped me recognize my true passion for field primatology. Thanks also to Leslie Williams, Giuseppe Vercellotti, Adam

Kolatorowicz, Tim Gocha, Megan Ingvoldstad, Lori Critcher, Dara Adams, Erin Kane,

Ashley Edes, and Noah Dunham for helping shape my graduate school experience. Thank you to the many friends I have made at UW-Whitewater as well. Special thanks to Ellie

Schemenauer for her mentorship. Thanks to Chandra Waring for being a friend and confidant. Thanks to Tracy Hawkins, Veronica Fruiht, Rachel Chaphalkar, and Courtney

Luedke for the time spent writing (and chatting) together.

Finally, thank you to my family for your constant encouragement, support, and guidance. Thank you so much to Karl and Janet Olson and the rest of the Olson clan. Our many dinners, game nights, and conversations have been a welcome reprieve from the academic grind. You keep me grounded and sane and I appreciate all that you have done for me along the way. To my niece and nephew, Avery and Thomas Horst, your smiles

vii and laughter are the light that has made this process a little easier. To my brothers-in-law,

John Horst and Stuart Jones, I appreciate all our conversations and the advice you have provided me along the way. To my sisters, Sara Horst and Emily Walz, I would not be where I am today without your constant support and guidance. Sara, you are a force to be reckoned with and going through this process has been easier because I know that you will always have my back. Emily, as a roommate, friend, and unofficial counselor, I know I could not have survived graduate school or the writing of this dissertation without you. To my parents, Nick and Becky Walz, it is difficult to sum the appreciation I have for all you have done for me over the years. From financial support to hours of editorial work on this document to dozens of hours of travel time so you could literally be by my side through this process, I am forever grateful. Without you, this dream would never have been a reality. Finally, to my husband, Jeff Olson, you are my rock and my voice of reason. You help me see the world in ways I never knew possible and I could never have made it to Tanzania and back without your unending support.

viii

Vita

2002...... Bryan High School

2006...... B.A. Zoology, Ohio Wesleyan University

2008...... M.A. Anthropology, The Ohio State

University

2008 to 2013 ...... Graduate Teaching Associate, Department

of Anthropology, The Ohio State University

2014 to present ...... Lecturer, Department of Sociology,

Criminology, and Anthropology; and

Department of Women’s Studies, University

of Wisconsin-Whitewater

Fields of Study

Major Field: Anthropology

Table of Contents

Abstract ...... ii

Dedication ...... iv

Acknowledgements ...... v

Vita ...... ix

List of Tables ...... xii

List of Figures ...... xvi

Chapter 1: Introduction ...... 1

Chapter 2: Methods ...... 15

Tables ...... 35

Figures...... 39

Chapter 3: Do female partner preferences influence copulations? ...... 41

Introduction ...... 41

Results ...... 49

Discussion ...... 57

Tables ...... 67

Figures...... 76

Chapter 4: Do female copulation calls encourage mate guarding from preferred partners? ...... 80

Introduction ...... 80

Results ...... 90

Discussion ...... 97

Tables ...... 108

Figures...... 114

Chapter 5: Do female olive baboons directly compete for access to mates? ...... 122

Introduction ...... 122

Results ...... 128

Discussion ...... 134

Tables ...... 142

Figures...... 143

Chapter 6: Conclusions ...... 148

References ...... 162

Appendix A ...... 178

Appendix B ...... 185

Appendix C ...... 188

List of Tables

Table 2.1: Demographic information and observation dates and hours for females followed during the course of the study...... 35

Table 2.2: Ethogram of behaviors recorded during focal follows...... 36

Table 2.3: Number of fecal samples collected for each focal female and their mean estradiol concentrations during and outside of ovulation...... 37

Table 2.4: Loadings from PCA for male characteristics...... 38

Table 2.5: Loadings from PCA for female characteristics...... 38

Table 3.1: Copulatory and aggression rates while females were ovulating, non-ovulatory, and in consorts for 89 female-male dyads...... 67

Table 3.2: Females and the percent of males with whom they interacted frequently, consorted, and “preferred.” ...... 70

Table 3.3: Results of partial correlation controlling for total hours observed and female characteristics and the percent of males with whom females interacted, consorted, and preferred...... 71

Table 3.4: Results of correlation between male characteristics and proportion of focal females with whom they interacted, consorted, and were classified as preferred...... 71

Table 3.5: Results of Mann-Whitney U comparing characteristics of preferred and non-preferred partners across all 89 dyads who interacted frequently during the study...... 72

Table 3.6: Reduced model testing influence of male characteristics on female partner preferences ...... 72 xii

Table 3.7: Generalized linear mixed models with AICc values that best explain whether a male contests consorts among olive baboon females ...... 72

Table 3.8: Generalized linear mixed models with AICc values that best explain whether a male goes on to consort with females whose consorts they contested...... 73

Table 3.9: Results of correlation comparing female rank and age to copulation rates during ovulatory periods, non-ovulatory periods, consorts, and complete copulations...... 73

Table 3.10: Results of correlation comparing male characteristics to copulation rates during ovulatory periods, non-ovulatory periods, consorts, and complete copulations...... 74

Table 3.11: Results of correlation comparing rates of aggression female received during ovulatory and non-ovulatory periods from males with female rank and age. ...74

Table 3.12: Results of correlation comparing male characteristics with rates of aggression males directed towards females during the ovulatory and non-ovulatory periods of focal females receiving the aggression...... 75

Table 3.13: Generalized linear mixed models with AICc values that best predict copulations among olive baboon females during different social and reproductive contexts...... 75

Table 4.1: 18 example predictions of the two competing adaptive hypotheses ...... 108

Table 4.2: Female identity, group, observation hours and copulation and copulation calling rates for the 19 study females...... 109

Table 4.3: Generalized linear mixed models with AICc values that best explain female copulation call production among olive baboon females...... 109

Table 4.4: Correlation between copulation call features and female rank and age...... 110

Table 4.5: Correlation between copulation call features and male characteristics...... 111

xiii

Table 4.6: Results from GLMM testing effects of male and female characteristics on a variety of call features across calls from 13 female olive baboons...... 112

Table 4.7: Generalized linear mixed models with AICc values that best explain whether a male contests consorts among olive baboon females ...... 113

Table 5.1: Generalized linear mixed models with AICc that best explain rates of aggression received and directed by cycling olive baboon females...... 142

Table 5.2: Generalized linear mixed models with AICc values that best explain rates of aggression directed by cycling olive baboon females in consorts...... 142

Table A.1: Information on female dominance rank derived from Elo-ratings ...... 178

Table A.2: Social distance values between female-female dyads in AC group in the 2012 study period ...... 181

Table A.3: Social distance values between female-female dyads in the BA group in the 2012 study period ...... 181

Table A.4: Social distance values between female-female dyads in the BA group in the 2014 study period ...... 182

Table A.5: Social distance values between female-female dyads in the DC group in the 2012 study period ...... 182

Table A.6: Social distance values between female-female dyads in the DC group in the 2014 study period ...... 183

Table A.7: Information on female dominance rank derived from Elo-ratings ...... 184

Table B.1: Male-Female Association Indices (AIs) for group AC ...... 185

Table B.2: Male-Female Association Indices (AIs) for group BA during the 2012 study period ...... 185

Table B.3: Male-Female Association Indices (AIs) for group BA during the 2014 study period ...... 185

xiv

Table B.4: Male-Female Association Indices (AIs) for group DC during the 2012 study period ...... 186

Table B.5: Male-Female Association Indices (AIs) for group DC during the 2014 study period ...... 186

Table B.6: Female partner preferences for Group (a) AC, (b) BA, (c) DC...... 187

Table C.1: Female-female dyads, kinship, and Affiliation Indices (derived from proximity and grooming interactions) during and outside of consorts ...... 188

List of Figures

Figure 2.1: Example of cyclical production of estradiol and corresponding swelling cycle for female HRF...... 39

Figure 2.2: Example of one female’s (UNK) Association Index (AI) with males in group AC...... 40

Figure 2.3: Female copulation call showing call duration, inhales, and exhales...... 40

Figure 3.1: Spearman correlation between male age and the proportion of focal females with whom males were observed interacting...... 76

Figure 3.2: Boxplots showing variation in relative association indices (pooling all individual female data) across males in each study group...... 77

Figure 3.3: Percent of times a consort was challenged by a preferred or non- preferred male and percent of times the challenge was successful ...... 79

Figure 4.1: Comparisons in three copulation calls of three savanna baboon species. ..114

Figure 4.2: Individual variation in calling during complete copulation in which the male ejaculated or incomplete copulations in which the male did not ejaculate...... 115

Figure 4.3: Individual variation in calling during contested and uncontested consorts ...... 115

Figure 4.4: Individual variation in calling following complete copulations in the presence or absence of any listener males...... 116

Figure 4.5: Spearman correlation to compare mean call unit rate and male sociality. .116

Figure 4.6: Differences in the number of units between calls expressed by females who are in their ovulatory phase and calls expressed by females outside the ovulatory window...... 117 xvi

Figure 4.7: Differences in the number of exhales between calls expressed by females who are in their ovulatory phase and calls expressed by females outside the ovulatory window...... 117

Figure 4.8: Comparison of the mean call duration (in seconds) between calls expressed while copulating with eschewed partners, neutral partners, and preferred partners...... 118

Figure 4.9: Comparison of the mean exhale rate (number per second) of copulation calls expressed with eschewed, neutral, and preferred male partners...... 118

Figure 4.10: Spearman correlation to compare calling frequency and number of days spent in consort...... 119

Figure 4.11: Copulation calling frequency versus average number of male partners per female...... 119

Figure 4.12: Comparison of the mean + SE dart distance (in meters) of females following complete copulations (in which a male ejaculated) and incomplete copulations...... 120

Figure 4.13: Mean dart distances (in meters) following completed copulations across females of different ages ...... 120

Figure 4.14: Comparison of the mean + SE dart distance (in meters) of females during copulations that took place during different phases of the cycle...... 121

Figure 5.1: Comparison of the mean rate (number per hour) of aggression (a) received and (b) directed among cycling females based on rank (with higher ranks being higher numbers...... 143

Figure 5.2: Comparison of the mean rate (number per hour) of aggression that focal cycling females (a) received and (b) directed towards other females of varying reproductive states...... 144

xvii

Figure 5.4: Comparison of the mean rates (number per hour) of aggression focal cycling females who were either ovulating or not and directed towards other females of varying reproductive states...... 146

Figure 5.5: Comparisons of mean Affiliation Indices while females are both guarded and not guarded among (a) nonkin and (b) kin ...... 147

xviii

CHAPTER 1: INTRODUCTION

The theory of sexual selection is an important explanatory framework for understanding the potential for specific morphological, physiological, and behavioral traits to evolve within one sex of a species to enhance their likelihood of successfully reproducing and thus increasing the fitness of individuals possessing such traits (Darwin

1871). Sexual selection theory is key to understanding how sexual conflict emerges as a result of differential investment in reproduction on the part of males and females in a species (Trivers 1972).

Primates express extended life histories and complex social behaviors compared to other animal taxa making it difficult to test mechanisms of sexual selection in our closest living relatives. Thus, biological anthropologists are still developing appropriate methods for understanding mechanisms of sexual selection. Particularly challenging is teasing apart female choice from sexual coercion and male-male competition in species with exaggerated sexual dimorphism like in many multimale, multifemale primates

(Setchell & Kappeler 2003; Palombit 2014). I aim to continue advancing this endeavor and specifically address proximate behavioral mechanisms, in particular those sexual and aggressive behaviors, that may increase or limit some females from accessing and successfully mating with potentially high-quality mating partners in olive baboons (Papio anubis). In this chapter, I introduce the interplay of the various sexual selection

1 mechanisms acting on nonhuman primates and outline how my study explores the impact of these mechanisms on female-female interactions, male-female interactions, copulatory rates, and copulation calls, a specific sexual signal that has been hypothesized to function in post-copulatory sexual selection. The questions guiding this research are as follows:

(1) What male characteristics, if any, increase the likelihood that female olive baboons will consistently solicit males around the time of ovulation? If females solicit certain males around ovulation, does this translate to higher copulatory success rates?

What role does male aggression play in influencing copulatory success among male- female dyads? I explore these questions in Chapter 3.

(2) Are copulation calls a potential post-copulatory mechanism of sexual selection in olive baboons as appears to be the case in other Papio species and primates in multimale, multifemale species more generally? Does more frequent calling influence the likelihood that females successfully copulate with preferred or otherwise “high-quality” partners? I address these questions in Chapter 4.

(3) Do female olive baboons directly compete for mates or compete in indirect ways, specifically for male “friends” or by suppressing reproduction in other females, as appears to be the case in other savanna baboons? This is the focus of Chapter 5.

This research has important implications for how we understand the complex interplay between male and female reproductive and competitive strategies and the sexual conflicts that occurred during the evolution of primates, including humans.

Sexual Selection and Sexual Conflict

The modern interpretation of intrasexual selection is derived from Darwin’s original description of the “sexual struggle” between members of the same sex to “drive away” their rivals, while intersexual selection is the struggle between members of the same sex to “excite” those of the opposite sex such that they will be chosen as a sexual partner (Darwin 1871). From his observations, Darwin suggested that males were generally the “competing” sex and females the “choosing” sex. Although it is now recognized that these “sex roles” may be reversed or relaxed depending on species life- history features as well as the social context in which organisms are imbedded (Small

1993; Setchell & Kappeler 2003), Darwin’s declaration that males generally compete for reproductive access to females and females choose males according to the traits they advertise continues to be the cornerstone of sexual selection theory. Advances in this theory have also made it evident that while coordination between the sexes is necessary to successfully reproduce, the reproductive objectives of males and females often do not

“harmonize perfectly” (Palmobit 2014, p. 237), which can result in evolved mechanisms of sexual conflict, especially male coercive tactics and female counterstrategies to male aggression (Trivers 1972; Smuts & Smuts 1993; Palombit 2014).

Bateman (1948) tested these phenomena empirically through experiments that demonstrated differential reproduction in the sexes of Drosophila. Three differences between the sexes were described: (1) male reproductive success is more variable than female reproductive success (2) the ability to attract a mate does not limit female reproductive success (3) female reproductive success increases very little with multiple

3 copulations (Bateman 1948; Trivers 1972). Gowaty (2012) recently demonstrated that because Bateman used marker mutations in offspring groups his samples likely were biased. Gowaty (2012) replicated Bateman’s experiments, and demonstrated that, similar to males, females are likely to seek multiple partners to ensure reproduction, which has been revolutionary in the study of sexual selection in diverse taxa (Gowaty 2012; see also

Scelza 2013). Although Bateman’s conclusions may have been flawed, his concern with differential investment in both gametes and offspring was revolutionary.

Until the 1970s it was relatively unclear why males were the more “competitive” sex and females the more “choosy” sex (Small 1993; Andersson 1994). Trivers (1972) considered the effects of anisogamous sex cells on these “sex roles.” Because males invest considerably less metabolic energy in production and maintenance of small, mobile gametes relative to female investment in large, immobile gametes (Bateman 1948;

Trivers 1972), males can more typically increase their reproductive success via finding and fertilizing as many females as possible, while females increase reproductive success through investments in their gametes and offspring. Trivers’ defined parental investment as, “any investment by the parent in an individual offspring that increases the offspring’s chance of surviving (and hence reproductive success) at the cost of the parent’s ability to invest in other offspring” (Trivers 1972, p. 139). He suggested that if one sex invests much more in offspring production and maintenance, then members of the opposite sex should compete amongst themselves to mate with the “investing” sex (Trivers 1972;

Krebs & Davies 1993). Among mammals, primates are characterized by especially slow life histories (Charnov & Berrigan 1993) and females invest considerable time and

4 energy in each offspring they conceive (Trivers 1972). Thus, due to their greater energetic investment in offspring through the processes of gestation, lactation, and postnatal care, females express higher parental investment and greater overall lifetime

“parental effort,” while males express greater lifetime “mating effort" (Krebs & Davies

1993). Males have greater potential fertility, and are under stronger selective pressures to compete for access to females (Trivers 1972; Krebs & Davies 1993; Andersson 1994).

Advances in the study of sexual selection, have centered around this parental investment paradigm and it is now recognized that, in addition to the classic forms that sexual selection takes in the Darwinian sense (e.g., inter- and intrasexual selection), sexual conflict is also a mechanism of sexual selection (Smuts & Smuts 1993; Clutton-

Brock & Parker 1995; Chapman et al. 2003; Palombit 2014) that, “emerges where

[reproductive] strategies of one sex impose fitness costs on the other sex” (Palombit

2014, p. 192), which can lead to an “evolutionary arms race” within a species between the sexes. In particular, it can lead to evolved strategies of harassment, intimidation, and mate-guarding among males to increase their rates of reproduction. Among females, coevolved strategies to counter the costs associated with male imposed strategies can include both behavioral and physical features. Examples include polygynandrous mating

(Wolff & McDonald 2004) as well as copulating more frequently or nearest to conception with dominant or other “preferred” individuals (Stumpf & Boesch 2005, 2006). Sexual signals, like sexual swellings, which generally increase in size around the time of ovulation can act as a “graded signal” of female fertility status (Nunn 1999). This may confuse the time of conception to some degree allowing females time to both confuse

5 paternity or choose among available males. Similarly, the production of copulation calls

(and other behaviors such as postcopulatory darts, or running away from a male immediately following copulation; O’Connell & Cowlishaw 1995) may act as a part of such a graded signal system (but see Engelhardt et al. 2012). However, assuming copulation calls vary on a finer scale then sexual swellings, they could function to increase opportunities for a female to continue a mating relationship or alternatively, because sound carries farther than visual cues, they may advertise to a wider audience when they are receptive and encourage mating with multiple males (Dixson 1998).

Throughout this dissertation, I specifically address how these sexual conflicts may be reflected in both male and female behavioral strategies in olive baboons at Gombe

Stream National Park.

Sexual Coercion

One of the primary ways sexual conflict manifest in primate species is the interaction between male sexual coercion and female reproductive strategies. In their seminal paper, Smuts & Smuts (1993) defined male sexual coercion as the “use by males of force, or threat of force that functions to increase the chances that a female will mate with him at a time when she is likely to be fertile, and to decrease the chances that she will mate with other males, at some cost to the female” (p 2-3). Competition for mates is intense in many primate species, and sexual coercion seems to be a by-product of selection for competitiveness that males also use to secure direct access to females. Smuts and Smuts’ theorized that males may use aggression towards fertile females in order to

6 immediately copulate with these females or copulate with them in the future. They presente numerous examples of male aggression they suggest function as coercion.

Infanticide, in which males may kill dependent offspring of females, is one such influential mechanism of coercion. This imposes obvious costs to females, and may increase the chance the males can mate with females because it induces female receptivity (Hrdy 1979). Clutton-Brock and Parker (1995) subsequently described three forms of coercion and reflected on the males of different species that use each type.

Forced copulations may occur when males use strength or speed to induce sexual intercourse. Sexual coercion can also occur as harassment, in which males frequently attempt to mate with females and it imposes a cost on the female. Intimidation and punishment is the final form that sexual aggression might take. Both clarified mechanisms by which male aggression may serve as coercion as well as why these strategies may evolve.

Since Smuts and Smuts published their seminal paper (1993) on male sexual coercion, it is now recognized that male coercive tactics influence the likelihood that females “preferentially” mate with specific males or exert choice within those species in which coercion is common. The threat of male coercion likely drives the tendency of females in many multimale primate species to mate polygynandrously during a given estrous cycle, instead of mating exclusively with preferred partners (e.g., red howler monkeys, Alouatta seniculus, Crockett & Sekulic 1984; long-tailed macaques, Macaca fascicularis, de Ruiter et al. 1994; chacma baboons, Papio ursinus, Palombit et al. 1997; diademed sifaka, Propithecus diadema ewardsi, Erhardt & Overdorf 1998; Japanese

7 macaques, Macaca fuscata yakui, Soltis et al. 2000). Mating with many males can increase paternity uncertainty across multiple partners and reduce the likelihood of male coercion, particularly in the form of infanticide (van Schaik et al. 2000; Clarke et al.

2009). Because of the considerable time and energy females invest in each offspring they conceive (Trivers 1972), infanticide is a significant threat to female reproductive success.

Strategies, like inciting paternity confusion that significantly reduce the likelihood of male coercion may outweigh the benefits of enacting choice on the basis of mate preferences (Small 1993; Heistermann et al. 2001).

Is There Room for Female Choice?

Although female primates are no longer viewed as passive, coy participants in male sexual escapades (Small 1984, 1993), as Darwin and early sexual selection researchers frequently suggested, it has been difficult to resolve how females can influence mating in cases where sexual conflict is high. However, promiscuous or polygynandrous mating does not necessarily have to prevent females from preferentially mating with certain males over others. Instead, females may behaviorally advertise their ovulatory cycles’ fertile phases to males who demonstrate preferred mate qualities (van

Schaik et al. 2004; Clarke et al. 2009). This can increase mating frequencies with preferred males and thus increase their paternity probabilities (Henzi et al. 2010). Subtly biasing paternity probability in this way can encourage males to provide females direct benefits in the form of offspring care, protection from conspecifics and predators, and access to resources, and thus male coercion may not preclude female choice even in

8 species expressing high rates of male aggression (e.g., chimpanzees, Pan troglodytes,

Stumpf & Boesch 2006, but see also Muller et al. 2011).

The hypothesis that females bias paternity towards preferred males by mating more frequently with them around the time of ovulation will help reconcile and explain, why, within a given species, males express both coercive strategies and provide care for infants (e.g., Papio spp. Palombit et al. 1997). Furthermore, this hypothesis can be applied to tests of both pre-copulatory and post-copulatory female choice mechanisms, and provides an appropriate explanatory framework for examining female preferences, especially at the proximate, behavioral level. I test this hypothesis by determining whether female olive baboons actively engage with, solicit, and resist copulations from specific males and produce copulation calls more frequently or whether calls differ in form with these males and whether such behaviors covary with conception likelihood

(Kappeler 2012).

Framework for this dissertation

In many mammals, behaviors associated with female proceptivity, or the active solicitation of a male through presenting behavior, approaching, and proximity maintenance (Beach 1976; Dixson 1998), are measures of the females’ hormonally motivated state and are exclusively expressed when she is most fertile (Young 1961;

Beach 1976). Serum concentrations of estradiol, the active form of estrogen, released in the blood peak just prior to ovulation and attach to estrogen receptors (ERs) in the brain

(Eisenfeld 1969; Blaustein 2009). This increase in estradiol binding likely mediates

9 proceptive behaviors. However, among most anthropoid primates still expressing cyclic production of estradiol during their fertile phase (Wallen & Zehr 2004), initiation of sexual behavior is not dependent on female sexual “motivation” as in other mammals, including prosimians (Pfaus 1999). Independence of hormones and sexual behavior facilitates sex outside the fertile period. Female motivation to participate in behaviors directly enhancing fitness by increasing the likelihood of conception may be distinct from female initiation and participation in sex for social purposes (e.g., reducing the risk of infanticide) (Wallen 1990; Zehr et al. 1998; Dixson 1998; Setchell & Kappeler 2003;

Wallen & Zehr 2004). While not limited to periods around peak fertility, proceptive and receptive behaviors corresponding with production of estradiol in particular peak concentrations, can proximally reflect motivation and ultimately function to increase the likelihood of conception (Stumpf & Boesch 2005, 2006). Males frequently receiving these kinds of behaviors around the time of likely conception could be considered targets of female choice (van Schaik et al. 2004; Stumpf & Boesch 2005). These were the criteria I use to assess female “preferences” in olive baboons, which I describe in more detail in Chapter 2 and Chapter 3. I also discuss the importance of considering whether males to whom females direct sexual advances towards might also be the males who are most likely to act aggressively towards females, and thus female behavioral choices may reflect a balance of fitness benefits and social benefits.

Primates in multimale, multifemale species often live in close association with their potential mating partners, and in some cases they have associated with these males for years. Thus, choice for an arbitrary genetic marker in the Fisherian sense (Fisher

1915) may be less likely than choice for a trait that is indicative of male fitness or ability to provide female with direct access to resources (Kappeler & van Schaik 2002). This may explain why, at least among cercopithecine primates, females appear to consistently choose to mate with dominant males (Engelhardt et al. 2006), who are most likely able to provide females protection and access to resources (Small 1989; Paul 2002). Choice for dominant males also may evolve because high social status may serve as an indicator of individual reproductive success. Because not all males are able to acquire dominant status in their life-time, those that do may reflect a form of “genetic superiority” that a female’s offspring could inherit. This might explain why female primates in species in which multiple partners are available at any given time frequently attempt to solicit dominant males during the fertile periods of their ovulatory cycles (reviewed by Setchell &

Kappeler 2003).

Although male reproductive success among cercopithecine primates is typically skewed towards high-ranking individuals (Alberts et al. 2003), this should not be taken as confirmation that these are the males with whom females prefer to mate. In part this is because these males may be more likely to be able to harass females into mating with them. Also, there may be other social makers of “mate quality” that are over-looked in female choice studies in favor of focusing on current male rank. For example, instead of biasing paternity towards high-ranking males, females may bias paternity towards young males that are likely to ascend the dominance hierarchy quickly, or already ascending the dominance hierarchy regardless of their age, given the benefits females could gain from these males in the future (Paul 2002; van Schaik et al. 2004; Rakhovskaya 2013). The

11 role of current rank versus “future” rank among males likely has significant impact on female mating preferences, especially in species typically characterized by large multimale, multifemale social structures. However, there are few well-habituated primate groups in which rank over time can be assessed. Thus, future rank is rarely incorporated into female choice studies, and in fact has only been demonstrated as a measure of male mate quality in studies using captive primates (e.g., rhesus macaques, Macaca mulatta,

Smith 1994). I consider this here and research correlates of male “rank trajectory” as well as current rank.

Females may also prefer “novel” males or males with shorter tenure in the social group, as has been recently demonstrated among Japanese macaques (Inoue & Takenaka

2008). Similar to the aversion female primates express for mating with close “childhood” associates, this preference for novelty may be a selected mechanism for incest avoidance or simply a mechanism to enhance genetic diversity (Paul 2002). I included multiple variables (e.g., current and future rank, age, tenure in group) that females may assess when attempting to both actively engage with specific males and copulate more frequently with those males, and expand on this in Chapters 2 and 3. One of my goals is to encourage primatologists to expand their definition of mate quality when testing female choice, and consider variables indicative of male fitness aside from current social rank that may be driving female mate preferences. I also describe why East African species of savanna baboons, especially populations like those at Gombe, which are consistently studied and have demographic records available, are particularly appropriate models for testing these mechanisms in Chapter 3.

Mechanisms of pre-copulatory choice have been studied extensively in a number of species (reviewed by Setchell & Kappeler 2003). However, it remains unclear whether females attempt to influence the likelihood that certain males maintain a short-term

“relationship” with them following a copulatory event, and thus behaviorally enact post- copulatory choice. Copulation calls may function as one such behavioral mechanism that allows females to maintain a consort with some males over others. In this way, these calls would encourage certain male partners to mate-guard, reduce the likelihood of copulating with other partners, and increase paternity estimates in preferred males (Maestripieri &

Roney 2005). This could be an especially effective mechanism in species where males effectively mate-guard females. I provide results of my test of copulation calls as a post- copulatory mechanism of sexual selection in olive baboons at Gombe Stream National

Park in Chapter 4 and compare them to other savanna baboons as well as their Asian counterparts, the macaques.

Next, I discuss the interaction of female-female competition and the expression of female mate preferences in olive baboons at Gombe in Chapter 5. Because reproductive success among female primates is tightly linked to their nutritional condition and investment in offspring, predation risk and food competition primarily govern the likelihood that females associate closely with kin and non-kin. However, food acquisition and predator avoidance are not the only pressures influencing female reproductive success. Recent research demonstrates that female primates also compete over mates, and inability to acquire viable partners can ultimately limit female reproductive success in a way that could contribute to variations in female fitness (reviewed by Cheney et al.

2012). This can occur even in species that express typical sex roles defined by the

Trivers-Bateman paradigm (baboons, Papio spp, Seyfarth 1978; stump-tailed macaques,

Macaca arctoides, Niemeyer & Anderson 1983; gorillas, Gorilla gorilla, Watts 1990; bonobos, Pan paniscus: Hohmann & Fruth 2003), particularly those in which females,

“face the highest costs of reproduction, OSRs are biased toward males, and males are the principal competitors” (Huchard & Cowlishaw 2011, p. 1003). For example, a recent study on wild chacma baboons (Papio ursinus) demonstrated that certain females were more likely to receive aggression from other females based on reproductive state and low-ranking females were especially susceptible to directed aggression from other females during consorts (Huchard & Cowlishaw 2011). These results have important implications for studies such as mine testing female mate preferences because female competitive strategies may influence individual ability to express preferences for specific male partners.

Finally, in chapter 6, I provide a summary and discussion of the results and conclusions derived from this study and demonstrate the interacting effects of male and female reproductive strategies and their impact on behavior and social relationships in olive baboons. I also demonstrate important differences in my study that differ from results on behavioral strategies of other savanna baboons. I describe important avenues for future research in the area of sexual conflict, sexual signaling, and female-female competition and address what might be done to fully assess the fitness consequences that these strategies may have for primates and humans.

CHAPTER 2: METHODS

Introduction

To address questions related to sexual conflict and interacting mechanisms of sexual selection in olive baboons outlined in Chapter 1, I used a combination of field and laboratory methods. In this Chapter, I describe my research site as well as the specific groups of baboons on which I focused. I also describe my techniques for collecting fecal samples for hormone analysis and behavioral data to assess stage in the ovulatory cycle and female partner “preferences,” copulation frequencies with specific partners, male- female dominance ranks, agonistic interactions, and attributes of copulation calls.

Study site and population:

Data were obtained at Gombe Stream National Park from March 2012 –

December 2012 (period 1) and March – June 2014 (period 2). Gombe is located 16 km north of Kigoma, Tanzania on the eastern shore of Lake Tanganyika. The park covers an area of approximately 52 km2 and is situated along the escarpment of the Great Rift

Valley. It consists of a series of ridges spanning from the lake to the escarpment. Because of this unique topography, the park comprises diverse landscapes and habitats with thick riverine forests characterizing the valleys and open grasslands spanning the ridges

(Clutton-Brock & Gillett 1979; Ransom 1981).

It is estimated that fourteen olive baboon (Papio anubis) groups inhabit Gombe

Stream National Park (DA Collins, Research Director, Jane Goodall Institute, pers. comm.). Individuals from nine of these groups have been consistently observed and studied since 1967, and thus are well habituated to human observers (Packer 1980;

Ransom 1981; Collins, unpubl. data). Major demographic events, including births, deaths, male transfers, and dates females initiate cycling, as well as matrilineal relationships have been recorded on these groups as part of the ongoing Gombe Baboon

Research Project (Collins, unpubl. data). I focused on three of these well-habituated troops (hereafter, AC, BA, DC).

Females were included as subjects if they had already had two regular sexual swelling cycles since weaning or since becoming mature by study onset. However, some females were dropped midway through the study if they conceived (identifiable based on the darkening of the perianal area) (Altmann 1973). If conception was suspected, I followed the female for at least 3 weeks to confirm, especially given that females can miscarry in the first few weeks. However, because a few females started cycling after study onset (and had at least two regular cycles) I was able to add some females for each female I dropped. Thus, over the course of the entire study, I followed 19 adult cycling females, following between 6 and 9 individuals at a time (Table 2.1). Fourteen of these females were followed in the 2012 study period and seven in 2014, with only two females studied in both periods.

Permissions

Research permits from both the Tanzanian Wildlife Research Institute (TAWIRI) and the Tanzanian Commission for Science and Technology (COSTECH) (No. 2012-12-

NA-2011-184 and No. 2013-336-ER-2011-184) were secured for this project. The

Internal Animal Care and Use Committee at The Ohio State University also approved this research (Protocol Numbers: 2012A00000006 and 2012A00000006-AR2).

Observational methods

During the first six weeks of the initial study period (March-April 2012), I learned to identify all adult baboons (with help from the Gombe field assistants), trained field assistants on sample collection techniques, and established an appropriate ethogram of behaviors on which to focus (Table 2.2).

With two field assistants, who aided in tracking and maintaining distance with female study subjects and collecting fecal samples (see below), I conducted a daily census and categorized reproductive status (i.e. cycling, pregnant, lactating) of all females in the group. During the remainder of the study, I collected behavioral data daily using multiple focal samples of cycling females in the group daily beginning at 0700 and ending at 1600 (Altmann 1974; Smuts 1985). I was in contact with baboon groups for approximately 2000 hours total. Typically, we switched between groups once per day or every other day. I sampled females in a random order and attempted to avoid re-sampling females until all focal individuals were sampled once (average n = 4 focal samples/day).

Neither randomization of group rotation nor female sampling could always be adhered to

17 due to the fecal collection protocol (see below). Focal periods were 90 minutes for females not in consort and 120 minutes for females in consort (similar to Smuts 1985).

Extra time when in consort allowed me to obtain more fecal samples and copulation call recordings around ovulation (See “Fecal Collection” below). I discontinued a follow if a female was lost for longer than 20 minutes. Across all cycle and reproductive phases, females were followed for a total of 1035.4 hrs (range = 27.65 – 111.28 hrs/female; mean

= 49.03 hrs/female) and 34 cycles (mean: 1.79/female; range: 1-4/female).

During focals, I used an all-occurrence sampling method (Altmann 1974) to record all interactions between females and other adults in the group (i.e. any time another individual was within 2 m of the focal) as well as the direction of the interaction

(i.e., whether the focal female was the active director or passive recipient of the behavior; see Table 2.2). This method allowed me to record a clear sequence of behaviors surrounding sexual interactions including identifying whether the male or female initiated sexual interactions, whether the female resisted or avoided the male, whether a copulation was complete (i.e., thrusts followed by a pause with evidence of ejaculate on the female’s sexual swelling), and the distance a female darted after a copulation (Semple et al. 2002).

I also opportunistically recorded agonistic interactions (e.g., threats and nasty behaviors defined in Table 2.2) and approach-retreat interactions between adults not involved in the focal sample. I used this ad libitum approach (Altmann 1974) because these behaviors are relatively uncommon yet are important for assessing dominance hierarchies (see Elo- ratings below). Proximity data were recorded every 5 minutes to determine the focal female’s nearest neighbor as well as individuals 0-2 m and 2-6 m away.

A consort was defined as a close spatial association between a male and female that lasted for at least 10 minutes, where individuals coordinated their movements and behaviors, and sexual activity was attempted (Bercovitch 1988). For females in consortship, I recorded whether males other than the consorting male were

“following”/contesting the consort. A contesting male is defined as one who follows the consorting pair for a period of at least 10 minutes, maintains a proximity of at least 5 meters, and coordinates behaviors in a similar way as the consorting pair (Danish &

Palombit 2014). During attempted consort takeovers, I recorded the identities of all males involved and whether the takeover was successful (i.e., a consort change-over occurs and the female has a new consort partner) (Bercovitch 1988).

Copulation Calls: Recordings

I used a Marantz PCM660 portable tape recorder and Sennheiser ME66 directional microphone to record female copulation calls during focals (Fischer et al.

2004). The identity of the caller and identity of the male were noted along with details surrounding the copulation (as above). When a male mounted a female, I approached as close to 3 m as possible, attempting to stand slightly in front and to the side of the female

(so as not to block her dart from the male), and directed the microphone towards the copulating pair. Male-female interactions during the 10-minutes immediately postcopulation were recorded (i.e., if copulation occurred at the end of a focal, data were collected for an additional 10 minutes or if no follow was underway, we followed her for

10 minutes).

Hormone Data and Reproductive Status

Fecal Collection and Processing

Hormones were extracted from fecal samples collected from target females using non-invasive techniques. My assistants and I maintained relatively close proximity (i.e. as close as 3 meters) with individual animals, allowing easy sample collection and ensuring we knew the producer of the sample. Using a targeted schedule, we collected approximately one fecal sample/3 days/cycling female and attempted to collect samples more frequently (attempting 1 sample/day or every other day) during the days surrounding maximal tumescence (following procedures similar to Higham et al. 2008;

Daspre et al. 2009). We collected a total of 612 samples (Table 2.3: mean = 32 samples/female; range = 16-68 samples/female).

I processed samples in the field using protocols developed specifically for baboon fecal hormone analysis (Beehner & Whitten 2004). After a focal individual defecated, I added approximately 0.5 grams of the fecal sample to 3 ml of a methanol/acetone solution (99.5% methanol; 8:2, measured using Gilson Pipetman Pipette, 1000µL) and homogenized the sample using a handheld biovortexer (Spectrum Chemical Mfg Corp). I extracted reproductive hormones using a double-filtration method 4-8 hours after samples were collected. During extractions, 2.5 ml of the fecal homogenate solution was filtered using polytetrafluoroethylene (PTFE) syringeless filters (0.2 µm; Whatman, Clifton, NJ).

Filters were washed with 0.7 ml of the methanol/acetone solution. The filtrate was diluted with 7 ml of distilled water and loaded onto primed octadecylsilane (C18) solid-phase

20 extraction cartridges (Waters Associates, Milford, MA). Cartridges were washed with aqueous methanol (99.5% methanol; 2:8) and stored in provisional desiccators (Whirl-

Pak bags filled with approximately 2 g of silica gel) for approximately 2 weeks.

Cartridges with extracted hormones were frozen after completely dry (freezers housed in the Gombe Ecosystem Health Project Laboratory). The remains of the solid fecal sample were left to dry in a desiccator and weighed approximately 2 weeks from the date of collection. Weights were recorded to the nearest hundredth of a gram.

Radioimmunoassay for Estradiol

Samples are thawed and shipped for radioimmunoassay (RIA) for estradiol (E2) at the University of Michigan’s Core Assay Facility (CAF). Although CAF conducted analyses, I worked at CAF (August 2014) assisting with the elution of samples and learning RIA techniques. For serum determinations of E2, CAF uses reagents from a

Pantex Direct 125I Estradiol RIA kit (catalogue 174M, Pantex, Santa Monica, CA), which have been validated for use of RIA of fecal estrogens and shown to reflect ovarian functioning in baboons (Beehner & Whitten 2004; Gesquiere et al. 2007). The primary antibody in this kit cross-reacts 100% with estradiol-17h, 5.6% with estrone, 2.63% with ethynylestradiol, 1.9% with a-estradiol and 0.68% with estriol (Pantex). Samples were validated for assay precision using both an inter- and intra-assay Coefficient of

Variability (CV). Inter-assay CVs of less than 15% are generally acceptable and intra- assay of CVs of less than 10% are generally acceptable (Parr, pers. comm.). My samples were acceptable for calculating E2 with inter-assay CVs of 8.24% and 9.65% (calculated

21 from measures of high and low concentrated quality controls) and an intra-assay CV of

4.76%. I standardized assays for differences in extraction volume and fecal weight and report individual values as ng hormone/g fecal dry weight (See Table 2.3 for mean E2 concentration during and outside of ovulatory periods for individual females).

Evaluation of Female Ovulatory Period

I graphically analyzed individual female hormone profiles to determine dates of ovulation for each individual female cycle and the corresponding fertile period of individual ovulatory cycles (for example: see Figure 2.1). For each individual cycle, the fecal E2 peak represents the approximate date of ovulation (Jeffcoate 1983; Higham et al.

2008). If samples were unavailable on a daily basis, then I incorporated the 1-2 days on either side of the peak concentration as ovulatory dates. Based on life spans of both sperm (three days) and egg (one day) (Behboodi et al. 1991; Wasser et al. 1994), the fertile phase is defined as the three days prior to the date of ovulation, the date of ovulation (or dates of ovulation) and one day following ovulation (Higham et al. 2008).

The mean fertile period across all females is 8 days. I conducted separate analyses during fertile and non-fertile periods for some comparisons.

I also assessed relative stage of the ovulatory cycle based on the size, color, and turgidity of sexual swellings daily (Domb & Pagel 2001). The following categories are used to classify sexual swellings of cycling females: flat (0) or menstruating

(corresponding to onset of cycle, no tumescence); inflated (1-2U) (perineal skin partly swollen, bright pink coloration); full (3) (tumescent, lustrous perineal tissue, absence of

22 fine wrinkles, light pink coloration); and deflated (1-2D) (partial swelling, multiple wrinkles, light coloration) (Daspre et al. 2009). Ovulation typically occurs about 1-4 days prior to detumescence (Shaikh et al. 1982; but see also Higham et al. 2008). Thus, as a further validation of the use of E2 as an indicator of ovulation in this group, I confirmed that peak fecal estradiol correlated with the approximate date of ovulation based on the visual cues (for example see: Figure 2.1).

Evaluation of Male and Female Dominance Hierarchies

I use approach-retreat interactions to assess dominance ranks among males

(Kitchen et al. 2003) and females. I assess dominance ranks and trajectories using the

Elo-rating method, a procedure originally developed for rating chess players (Elo 1978) but now accepted as a useful tool for assessing dominance hierarchies in animal social systems, particularly nonhuman primates (Neumann et al. 2011). It is a particularly robust analytical tool for groups in which some individuals are never observed interacting or there are few observed interactions generally, which occurred among males and females in all three of the troops observed, or in which there are 5 or fewer individuals vying for dominance, as was the case with males in BA group during the first study period.

Within this rank-modeling system, individual ratings increase or decrease based on the outcome of the contest and the likelihood of a particular outcome. For example, if one individual wins a contest against another individual with a lower accumulated Elo- rating, each individual’s scores will increase and decrease marginally. However, if an

23 individual wins a contest against another individual with a greater accumulated Elo-rating

(i.e., a higher-ranking individual) each individuals’ scores will change more substantially.

Thus, this system relies on assessing the degree of conformity to the likelihood of winning or losing contests based on accumulated scores over time.

Using the EloRating package (version 0.43: Neumann & Kulik 2014) in R

(version 3.1.3), I established Elo-ratings for each month and calculated the mean value across each study period (2012 and 2014) for males and females in each of the three troops. At the start of the study period, initial ratings for individuals were valued at 1000.

I set the number of points awarded or lost during individual contests (k) at 100 and these points were weighted using the probability of winning or losing the encounter as described above and following numbers used in Neumann et al (2011) and Johnson et al.

(2014). Individuals were rank ordered based on their mean Elo-ratings for each study period and then I assigned Elo-ratings an ordinal rank (i.e, 1, 2, 3 with higher numbers indicating higher rank) (Tables A.1 and A.7). To compare across individuals in troops of variable sizes and sex ratios, relative ranks of males and females within each troop were evaluated as:

(푁푥 − 푂푅푦) ∗ 100 (푁푥 − 1) where Nx is the number of adult males or females in a troop and ORy is individual y’s ordinal rank, such that high relative rank reflects high rank in the dominance hierarchy

(Tables A.1 and A.7). Using the traj.elo function in R’s EloRating package, I also assessed male rank trajectory for each individual during each study period. In this program, values further from zero indicate an individual is rising or falling in rank over a 24 given time. Positive values indicate an individual is rising in rank while negative values indicate an individual is decreasing (Table A.7).

Finally, the Elo-rating method can be used to determine social distance between females within a social group based on the difference in Elo-ratings of two individuals.

This technique has been used to investigate the strength of social bonds among female geladas (Theropithecus gelada) (Johnson et al. 2014). This is an important new measure of sociality in primate societies because it accounts for differences in social positions between two individuals who may represent similar rank positions in different troops

(e.g., the alpha female may be “closer” to the beta female in troop A than alpha and beta females in troop B) or at different positions in one troop (e.g., female in rank position three may be closer to female in rank one than female six to female four even though females three and six are both two rank positions away from females one and four respectively). In Chapter 5, I use social distance for each female-female dyad based on their mean difference in Elo-rating between two females (Johnson et al. 2014) (Tables

A.2-A.6). All rank information for males and females including rank trajectory and social distance values can be found in Appendix A.

Male-Female Associations and Female Mate Preferences

Female olive baboons interact with many males when cycling. For example,

Bercovitch (1987) reported that female olive baboons engage in consorts with up to 38-

75% of males in a group outside the fertility window and 50-100% of males in a group during fertile windows (which he defined as the days surrounding swelling

25 detumescence; Bercovitch 1987). Bercovitch reported females spent 66% of their fertile days not in a consort, but still are approached by multiple males and that when in consort, follower males were common. Up to 45% of consortships had a male turnover

(Bercovitch 1987; 1989). However, even though a female may have many males near her during her time of conception, she does not necessarily actively interact with (i.e., by approaching, grooming, or acting proceptively) or resist each of them. Thus, to determine partner “preference,” I identified whether each focal female singled out some males over others during her fertile phase.

Specifically, I developed a focal-centric association index (AI) based on a similar one developed by Girard-Buttoz and colleagues (2014), who combine total approach ratios and total groom ratios to determine male-female bond strength. However, I also incorporated measures of female proceptivity and resistance as these have been demonstrated to be important measures of female mate choice in primates (Bercovitch

1995; Stumpf & Boesch 2005, 2006; Engelhardt et al. 2006). I calculated an AI for each male-female dyads during ovulatory days:

푇퐴푅푖푗 + 푇퐺푅푖푗 + 푃푅푖푗 − 푅푅푖푗

4 where TARij is the approach ratio of female i to male j relative to all males in the troop;

TGRij is the total grooming ratio of female i with male j relative to all males in the troop;

PRij is the proceptivity ratio of female i with male j relative to all males; and RRij is the resistance ratio of female i with male j relative to all males. Next, I describe the different components of this equation and then I provide details on criteria used to categorize males.

Presentations that were unsolicited (e.g., did not involve a male approaching or touching a female) and occurred while females were in estrus were classified as proceptive behaviors (Smuts 1985; Stumpf and Boesch 2005). Female “presenting” behaviors involve displaying the perianal area to a male for inspection and these behaviors range in style and form from leaning forward and looking at a male between the legs (i.e., “peek-a-boo”) to lifting one back leg back towards the male. Resistance behaviors include darting from a male without copulating first (i.e., not a copulatory dart:

Semple et al. 2002) and avoiding a male (Ransom 1981; Smuts 1985; Stumpf and Boesch

2005; Bailey et al. 2015). Proceptivity (PR) and resistance ratios (RR) were calculated as:

푃푖푗 푅푖푗 ; 푃푖푥+푅푖푥 푃푖푥+푅푖푥 where Pij and Rij are the number of times female i was proceptive (P) or resistant (R) during the ovulatory period with male j relative to the times female i was proceptive and resistant to all males (Pix+Rix) during the ovulatory period. The denominator for both in this case is a summation of proceptivity and resistance measures because each measure was rare enough that, if kept singly, there would have been “0”s in the denominator. This is especially true given that I defined proceptivity here as “unsolicited” presents and female resistance is relatively rare in this group (Bailey et al. 2015).

Total approach ratios (TAR) were calculated as:

퐴푖푗

퐴푖푥 where Aij is the number of times female i approached male j relative to the number of times female i approached all males (Aix) (Girard-Buttoz et al. 2014).

Total grooming ratios (TGR) were calculated as:

퐺푖푗

퐺푖푥 where Gij is the grooming time between female i and male j relative to the grooming time of female i with all males (Gix) (Girard-Buttoz et al. 2014).

Results of these overall AI calculations are summarized in Tables B.1-B.5 in

Appendix B. For each dyad, I used male-female AIs during ovulatory periods to determine those males with whom females actively engaged. Following methods of

Stumpf and Boesch (2005), for each female, I determined average AIs across all males in the group and determined those who fell 25% above the mean and those who fell 25% below the mean. Hereafter, to be consistent with previous literature, I refer to those in the upper 25% as “preferred” and those in the lower 25% as “eschewed.” Males who were neither eschewed nor preferred males are “neutral” (See Example: Figure 2.2). In some analysis neutral and preferred males are lumped as “non-preferred” (See Chapter 3). This method allows all males to be categorized and also allows males to be classified in multiple categories if they were preferred, eschewed, or neutral to different females

(Table B.6). However, it is important to note that male competition might prevent some males from getting near females around the fertile period and thus might restrict options for female “choice.” Therefore, high AI scores do not mean that these are the only males with whom a female would have “preferred” to interact (Small 1993).

To compare across males within study groups and establish whether there are consistent patterns in female-partner preferences, I also calculated a relative association index for each male-female dyad:

퐴퐼푥 − x̅퐴퐼푥푖 where AIx is the Association Index of male x and x̅ AIxi is the mean Association Index across all males for female i (Stumpf and Boesch 2005).

Copulation Rates and Male-Female Aggression

In Chapter 3, I compare aggression and copulation rates among dyads based on whether the female was fertile or not and whether the dyad was in a consort (see Chapter

3). I restricted analyses to the subset of dyads that interacted frequently, defined as dyad that had at least 10 summed approaches and leaves during the various contexts (Soltis et al. 1997). I used this criteria (interact > 10 times) to test whether males and females who interacted significantly during all cycling days.

Copulation rates and rates of aggression received during each context (i.e., in a consort or not in a consort, while ovulating or not) were calculated as:

퐶푂푃푖푗 퐴퐺퐺푖푗 ; 푇푖푚푒푖 푇푖푚푒푖 where COPij are the number of copulations that occurred between female i and male j (on all ovulatory days for example) and AGGij are the number of times male j directed aggression towards female i (on all ovulatory days) (Kaburu & Newton-Fisher 2015).

Timei is the observation minutes for each female in each context.

Completed copulation ratios were calculated as:

퐶퐶푂푃푖푗

퐶푂푃푖푗

29 where CCOPij are the number of completed copulations that occurred between female i and male j (in a given context) and COPij are the total number of copulations that occurred between that male and female (in a given context such as in a consort vs. unguarded, fertile vs. unfertile, etc).

Copulation Calls

In Chapter 4, I analyzed all copulations in which it was clear whether a female called and whether a male ejaculated. I restricted analyses to only those females who engaged in at least 10 copulations over the course of the study (following Nikitopoulos et al. 2004). For each female, I calculated the proportion of calls that occurred during complete copulations and incomplete copulations in each context (when fertile vs. not fertile, when in consort vs. unguarded etc; see Chapter 4).

Calls were analyzed for overall call duration, number of inhales and exhales (i.e., the individual “units”), number and rate of total units/s per call (Semple et al. 2002), and exhale rate (Engelhardt et al. 2012; Figure 2.3) using Pratt and Raven bioacoustic software. Based on sound quality it was not possible to analyze calls for frequency measures.

Female-Female Affiliations

In Chapter 5, along with calculating social distance, for each focal female and within each reproductive context, I also calculated a female-female affiliation index

(modified from Johnson et al. 2014) between focal and other females in the group. For these calculations, I used ratios of grooming and proximity as follows:

퐺 푁 푖푘 + 푖푘 퐺푖푥 푁푖푥 2 where Gik is the grooming time between females i and k relative to the grooming time with all females in the troop (Gix); and Nik is the number of times female i was the nearest neighbor of female k during scan samples relative to all occurrences of females as the nearest neighbor to female i (Nix) (modified from Johnson et al. 2014). Results from these calculations are available in Table C.1 in Appendix C.

For the purposes of this analysis, female kin relationships were assessed using the long-term demographic data record housed at Gombe Stream National Park. Immediate kin were defined as mothers, daughters, and female siblings who share a mother because studies suggest that kin selection benefits significantly decline beyond these relationships

(Hamilton 1964) and DNA profiles were not available to assess paternal relationships in this population.

Male and Female Characteristics

Individual ages, group tenure (males), and number of offspring birthed (females) were copied from demographic records housed at Gombe as part of the ongoing Gombe

Baboon Research Project (Collins, unpubl. data), and rank information were taken from

Elo-ratings as described above. However, in some cases when there were too many variables to incorporate into Generalized Linear Mixed Models (GLMM; described below), I reduced these characteristics to one or two components so as not to lose too 31 many degrees of freedom in the models. To do this, I ran a factor reducing Principal

Components Analysis (PCA) separately for males and for females and accepted components for each of these analyses if the Eigen value exceeded 1.00. Resulting principle components were direct-oblimin rotated (Soltis 1999a).

For males, two principal components were retained in the PCA, together accounting for 70.1% of the variance in the data set. Table 2.4 lists the loadings for each of the original characteristics on the two components, which I labeled as “Male Sociality” and “Male Experience.” Male Sociality consisted of characteristics associated with male social rank, including current relative rank and rank trajectory, based on Elo-ratings.

High values indicated males of high rank who were ascending the hierarchy. Male

Experience consisted of characteristics associated with male familiarity with members of the group, including their tenure in the group based on approximate number of months in the group and age. High values indicated older males with longer group tenures.

For females, two components were also retained in the PCA, together accounting for 92.9% of the variance of the data set. Table 2.5 lists the loadings for each of the original characteristics on the two components, which I labeled as “Female Experience” and “Female Sociality.” Component 1 (Female Experience) consisted of characteristics associated with a maternal history, including the number of live offspring birthed (i.e., parity), and age. High values indicated older females with multiple births. Component 2

(Female Sociality) consisted of only female relative rank so therefore I just used rank in all analyses.

Statistical Analyses

Because my data typically were not normally distributed (determined using a

Kruskal-Wallis test), I primarily used non-parametric statistics. I used Mann-Whitney tests for between-female comparisons and Wilcoxon signed-ranks tests for within-female comparisons (Kaburu & Newton-Fisher 2015). I used Spearman’s rank correlation coefficient to correlate dyadic interaction rates and copulation calling rates during different social and reproductive contexts (e.g., during guarded or non-guarded consorts)

(Engelhardt et al. 2012) with social (rank, rank trajectory) and experience variables (age, tenure). Analyses were two-tailed or one-tailed depending on whether previous research led to directional hypotheses (Gaines & Rice 1990; but see Lombardi & Hurlbert 2009) with α = 0.05.

Given that both individual males and females are represented in multiple dyads, these dyads are not completely independent. Therefore, I also used generalized linear mixed models (GLMM) to assess variation in dyadic interactions according to reproductive and social contexts. GLMM is a valuable tool for these data because I could control for and assess the influence of both random effects, particularly individuals, who contributed multiple points in each data set I considered, as well as the relative influence of fixed effects, like male and female sociality (rank or rank trajectory) and experience

(tenure in group, age, number of offspring birthed). In the case of data sets in which the response variable were normally distributed or could be transformed accordingly, linear mixed models (LMM) were utilized.

In GLMM, the response variable was typically either a binary variable (e.g., “call produced” or “no call produced”) (Cheney et al. 2010) and thus I fit the model to a logistic regression, or was a count of behaviors occurring over a specific period of time

(e.g., number of times a male directed aggression towards a focal female). These measures included zeroes, thus I fit the model to a Poisson distribution with log link functions (Kaburu & Newton-Fisher 2015). In Chapter 5, to allow for comparable results on possible influencers of female-female aggression in olive baboons with results on chacma baboons (Papio ursinus), I primarily followed the methods of Huchard and

Cowlishaw (2011). Here, GLMM were used to assess whether the proportion of cycling females predicted rates of aggression exchanged as well as significant predictors of aggressive interactions exchanged during focal follows.

For each analysis, I ran full models first, including all variables of interest. I then conducted a model selection procedure to identify the model that best predicted the dependent variable of interest. This process follows Burnham and colleagues (2011), where separate models with differing combinations of predictors are ranked on the basis of their corrected Akaike information criteria (AICc) values. AICc values correct for small sample sizes. Best fit models were those with the lowest AICc values (Burnham et al.

2011; Kaburu & Newton-Fisher 2015). Specific model structures are described in each of the following Chapters.

TABLES

Table 2.1: Demographic information and observation dates and hours for females followed during the course of the study.

Troop Short Birth Number of Dates Followed Total identifier offspring (Month/Year) Hours AC UVA 10/24/1993 6 08/2012-12/2012 30.93 UBG 6/19/1998 5 06/2012-07/2012 28.68 UNK 1/12/2006 0 06/2012-12/2012 61.38 ULY 4/18/2006 0 07/2012-/12/2012 46.17 YLT 6/2/2005 0 06/2012-07/2012 31.55 BA WTW 1/1/1995 5 05/2012-07/2012 47.67 WLD 1/8/2001 2 05/2012-06/2012 27.65 WDF 1/6/2001 0 05/2012-12/2012 111.28 WGR 1/27/2005 1 03/2014-05/2014 34.97 AJA 9/27/1995 4 04/2014-06/2014 42.15 AKA 4/11/1997 1 04/2014-06/2014 49.68 DC HRF 1/23/2006 0 08/2012-12/2012 52.00 HRT 7/10/2004 1 08/2012-11/2012 38.98 HON 1/30/2007 0 08/2012-12/2012 54.10 HRS 12/11/2003 2 08/2012-12/2012 90.29 04/2014-06/2012 SAS 9/13/2005 0 08/2012-12/2012 83.62 03/2014-05/2014 HAU 8/25/2002 3 03/2014-06/2014 79.62 SPR 1/2/2004 1 03/2014-06/2014 72.87 SER 5/24/2005 1 03/2014-05/2014 54.50

Table 2.2: Ethogram of behaviors recorded during focal follows.

CODE BEHAVIOR carry another’s baby (typically friendly behavior but can also be by male during an B aggressive interaction with another male. Detailed notes will be taken in this case). friendly/affiliative (when possible, notes will be taken to describe this behavior such as: inspect/sniff other’s mouth, lip to lip/”kiss”, touch gently but no grooming, hug, F males ‘chew’ on females, lay/rest in contact, males often greet one another by touching each other’s rumps or genitals while producing a fluttering vocalization) N nasty (chase, bite, slap, punch, hold down) T threat (flash eyebrows, chin thrust, sweep ground, stiff arm bounce, shake branch) S submissive (fear grimace, cower, lean away but do not leave area) D supplant/displace A approach (one animals moves to within 2m of another) X stay near L leave (one animals moves greater than 2m away from a close neighbor) W wish/present for groom G Groom Z end groom P present (rear) I inspect (rear) AV avoid (male-female; female moves in the opposite direction of an approaching male) M mount but no intromission (can be male-male and female-female) C copulate (information on intromission and complete or incomplete) R Postcopulation dart/run (also note the distance) grunts (short, quiet, low frequency calls; often done as one animal approaches VG another, when attempting to touch an infant and before moving into an open area) roar grunt (a loud call produced by males only, includes a sustained roar followed by VR 3-5 loud grunt-type vocalizations) VT threat grunts (distinctively short and louder, often given rapid sequences) VW wahoos (loud, low frequency call that can function as a contact call) VS screams (high pitched, noisy, protracted call) VC copulation calls (adult females’ distinctive long call followed by several loud grunts) VA alarm call (loud single or double bark calls) lost/contact call (very similar to alarm calls but acoustically noisier and given with VL less repetition and in different context than alarm calls) O other/unknown behavior Q indicates notes or ad libidum data 1 Active (action done by focal animal) 2 Passive (action done to focal animal)

Table 2.3: Number of fecal samples collected for each focal female and their mean estradiol concentrations during and outside of ovulation.

No. Samples Female Mean E2 (ng/g): Mean E2 (ng/g): Collected Ovulatory Nonovulatory AJA 16 4,033.29 1,266.45 AKA 22 4,133.21 1,813.65 HRS 40 7,557.66 2,858.52 HRT 29 5,809.04 3,118.52 HRF 36 7,754.19 3,367.11 HAU 35 5,861.20 1,264.38 HON 39 4,436.50 2,308.44 SAS 44 7,583.64 2,344.22 SER 17 9,398.75 2,465.96 SPR 26 5,567.82 1,735.48 UBG 28 3,635.38 2,357.47 ULY 36 4,601.60 2,046.48 UNK 44 3,873.31 2,038.82 UVA 25 2,588.72 1,363.85 WLD 17 3,455.26 1,190.35 WGR 19 3,892.70 1,496.96 WTW 44 1,562.58 567.58 WDF 68 3,397.51 1,715.97 YLT 27 5,591.20 1,617.49

Table 2.4: Loadings from PCA for male characteristics.

VARIABLES Male Sociality Male Experience Relative Rank 0.829 0.166 Trajectory 0.846 -0.209 Tenure 0.429 0.563 Age -0.152 0.886

Variance Explained 43.8 26.3

Table 2.5: Loadings from PCA for female characteristics.

VARIABLES Female Experience Female Sociality Relative Rank 0.001 0.998 Age 0.949 -0.104 Number of offspring 0.932 0.110

Variance Explained 59.4 33.6

FIGURES

20,000

18,000

16,000 14,000 12,000

10,000

8,000 6,000

4,000

E2 Concentrations (ng/ hormone/g feces) hormone/g (ng/ Concentrations E2 2,000

Sexual Swelling Size Category

Figure 2.1: Example of cyclical production of estradiol and corresponding swelling cycle for female HRF. Shaded areas correspond to the fertile or ovulatory period of the cycle.

UNK Partner Preferences 0.25 O 0.2

0.15

x O x x 0.1

Association Index 0.05

0 MEAN AMX AVZ GBS LRA SHT SYR Males

Figure 2.2: Example of one female’s (UNK) Association Index (AI) with males in group AC. The black, dotted line represents the value that is 25% above the mean. Males who have AIs with this female that fall above the black line represent “preferred” partners, which are also marked on the figure with a “O”. The gray, dashed line represents the value that is 25% below the mean. Males who have AIs below the gray line represent eschewed partners, which are also marked with an “x“.

Exhale Unit Inhale Unit

Call duration

Figure 2.3: Female copulation call showing call duration, inhales, and exhales.

CHAPTER 3: DO FEMALE PARTNER PREFERENCES INFLUENCE COPULATIONS?

Introduction

Females frequently mate with multiple males in multi-male, multi-female primate societies (reviewed by van Schaik et al. 2004). Although direct female choice for male quality is one potential explanation, mating with multiple males can provide indirect benefits such as permitting cryptic female choice through sperm competition or simply ensuring insemination (Small 1988; Møller & Birkhead 1989; Wolff & Macdonald 2004).

Polygynandrous mating may also function to dissuade infanticide because copulating with a male at least once during a female’s conceptive cycle means there is a chance he has sired her offspring and he should not kill her infant (van Schaik et al. 2000; van

Schaik et al. 2004; Clarke et al. 2009). Furthermore, because males may directly coerce females into mating via threats, aggressive attacks, and forced copulations (Smuts &

Smuts 1993), polygynandrous mating may function to appease sexually coercive males and allow females to avoid injury (Swedell & Schreier 2009). Thus, it is becoming increasingly clear that male-male competition and male ability to monopolize copulations with females can seemingly confound female mate preferences and potentially reduce female ability to exert preferences. Dominant males may simply be more likely to monopolize matings with females rather than females exerting strong preferences for

41 these males (e.g., rhesus macaques, Macaca mulatta: Manson 1992; Japanese macaques,

Macaca fuscata: Soltis et al. 2001; chimpanzees, Pan troglodytes: Muller et al. 2011).

Nevertheless, although subtle and potentially difficult to identify in societies in which male coercion and competition is common, it is hypothesized that females may still be able to influence mating patterns (van Schaik et al. 2000; van Schaik et al. 2004;

Clarke et al. 2009). For example, by behaviorally advertising the fertile phase of their ovulatory cycles to males who demonstrate desirable qualities, females can increase mating frequencies with preferred males and increase these males’ paternity estimates

(van Schaik et al. 2004; Clarke et al. 2009; Henzi et al. 2010). If these males are preferred based on direct or indirect (e.g., genetic) benefits they provide, then females could improve their fitness. Furthermore, such a strategy can be reconciled with other female counterstrategies thought to be common in multimale, multimale societies with high sexual coercion (see above). For example, females may still mate with several males throughout their ovulatory cycle to augment paternity estimates across multiple males, but mate with preferred partners most often, especially during the days surrounding ovulation (van Schaik et al. 2004). Males more “confident” in their probability of siring offspring should not only avoid infanticidal behavior, but also actively protect females and their offspring in the future, which offers significant fitness benefits in terms of increasing the likelihood infants survive vulnerable developmental periods (Palombit

2000). In this way, female behavior may function in a similar way as sexual swellings, which, according to Nunn’s “graded-signal hypothesis” (1999) serve as a sexual signal that both confuses and confirms paternity in males (van Schaik et al. 2004; Clarke et al.

2009). Thus, male strategies may not preclude female choice, even in species characterized by extreme sexual dimorphism, male dominance, and aggression (Stumpf

& Boesch 2005).

Despite the power of this theoretical framework, the strategies females may use to bias paternity towards particular mates in non-seasonally breeding primates living in multimale societies have been systematically tested in the wild in only a few species (e.g. chimpanzees: Matsumoto-Oda 1999; Stumpf & Boesch 2005, 2006; orangutans, Pongo pygmaeus: Stumpf et al. 2008). Data from these species have provided inconclusive results on whether females can copulate more frequently with specific males as the likelihood of conception increases. For example, among the Taï forest chimpanzees, females mate more frequently with high-ranking males as conception probability increases (Stumpf & Boesch 2005, 2006). Stumpf and Boesch (2005, 2006) suggest this may be an attempt on the part of females to bias paternity towards these high-quality males to protect against infanticide. However, further investigations into this relationship indicate that, among chimpanzees, “preferred” male partners, and those in which females mate more frequently with during the periovulatory period, are also those who are particularly aggressive towards females over their entire ovulatory cycle (Muller et al.

2011). Thus, it may be that females mate more frequently with specific males out of

“fear” or concession rather than preference for these males, and that male coercive strategies still limit female ability to express mate preferences in chimpanzees (Muller &

Wrangham 2009; Muller et al. 2011). Chimpanzees may be somewhat unique in that males are particularly aggressive towards females (Muller & Wrangham 2009) and bonds

43 between males and females are relatively weak (Machanda et al. 2013), with males providing females few direct benefits in the way of protection and offspring care.

Unlike chimpanzees, savanna baboons are an ideal model for testing whether females bias paternity toward potentially protective males in the context of a multimale social system (Palombit 2000). Although polygynandrous mating is the norm among savanna baboons (Seyfarth 1978a; Smuts 1985; Bercovitch 1995), males and females also form close, nonsexual bonds, termed “friendships” throughout the literature (Strum 1975;

Altmann 1980; Smuts 1985; Palombit et al. 1997). Both parties benefit from these dyadic associations through increased grooming, access to resources, and predator avoidance

(Altmann 1980; Smuts 1985). Most significantly, males protect their friends and their friends’ offspring from conspecifics and predators (Ransom & Ransom 1971; Packer

1980; Smuts 1985; Palombit et al. 1997), something that is rarely seen outside of a mating context in any other species besides humans (Palombit 2009). Paternity data in chacma (Moscovice et al. 2010) and yellow baboons (Buchan et al. 2003) and endocrinological data in olive baboons (Shur 2008) suggest that males preferentially form friendships with the mothers of their offspring. Furthermore, baboons monitor relationships among group members (e.g., Kitchen et al. 2005; Crockford et al. 2007), and conspecifics are less likely to harass females when the victim’s male friend is present

(Smuts 1985; Lemasson et al. 2008). Therefore, although male baboons mate-guard females during prolonged consortships, females may still be able benefit from manipulating “confidence” in paternity (and corresponding future protective tendencies)

44 in select males by demonstrating affiliative and proceptive behaviors toward them as well as mate with them more frequently than others.

The East African savanna baboons (olive baboon: Papio anubis; yellow baboons:

Papio cynocephalus ) in particular offer an interesting opportunity to test female tendencies to bias paternity towards specific males because females in these species have more opportunities to play a significant role in consortship formation and maintenance

(Ransom 1981; Smuts 1985). East African baboons are characterized by intense male competition and dominance hierarchies (Strum 1975; Berenstain & Wade 1983; Smuts

1985). However, compared to their South African relative the chacma baboon (Papio ursinus) (Bulger 1993; Weingrill et al. 2000), male dominance rank does not necessarily govern access to reproductively capable females or consort formation among the East

African baboons (Strum 1982; Smuts 1985). For example, in their work on yellow baboons, Alberts and colleagues (2003) reported that males are less constrained by the priority-of-access (POA) model (Altmann 1962) that typically explains male rank and its relationship to reproductive success in chacma baboons (Bulger 1993). This yellow baboon “queue-jumping” behavior (Alberts et al. 2003) also seems the rule in olive baboons because dominance rank is not the only predictor of reproductive success

(Packer 1979b, Bercovitch 1987, Danish & Palombit 2014). The absence of POA is likely facilitated by the fact that males of all ranks and ages follow consorting pairs for several days in both species, with some of these males actively contesting with the consorting male for access to his female partner (Hall & DeVore 1965; Bercovitch 1988; Danish &

Palombit 2014). Conversely, such behaviors are limited to acute consort takeovers,

45 always by higher-ranking males in chacma baboons (Bulger 1993). Additionally, although males frequently initiate consortships, female olive and yellow baboons may refuse males’ displays or “hide” from their consorting partners (Ransom 1981; Smuts

1985; Bercovitch 1995), whereas such female resistance has not been seen in chacma baboons (Kitchen pers. comm.). Furthermore, although consorting males will sometimes herd females away from other males in all species, female olive baboons can still approach a variety of males even when in consort (Bercovitch 1995). Thus, females can actively “encourage” consort challenges (e.g., by approaching and acting proceptively towards non-consorting males), thereby potentially minimizing likelihood of a protracted consortship with a less preferred consort partner (Rasmussen 1980). Perhaps some of the consort turnover unrelated to rank seen in East African species is because males are be more likely to concede power when females show consistent preference for specific males (Seyfarth 1978a; Smuts 1985), which is unlikely to happen in chacma baboons given the strong effects of POA (Bulger 1993).

In a comprehensive analysis of female olive baboon behavior surrounding consortships, Bercovitch (1995) demonstrated that female’s use resistance and proximity maintenance to mate with preferred partners rather than mate promiscuously. However,

Bercovitch only relied on female presenting and grooming behavior during nonconsort days leading up to ovulation, when in fact behavior should be vastly different during the ovulatory window than during cycling but not fertile days. Additionally, Bercovitch’s study was conducted from 1979-1981 when it was typical that the fertility window was distinguished from other phases of the cycle based only on sexual swelling size

46 categories rather than confirmation based on either photographic (Rigaill et al. 2013) or endocrinological data (e.g., Higham et al. 2008). Bercovitch also only considered the impact that male-female aggression has immediately surrounding copulation rather than consider the impact that consistent directed aggression throughout the ovulatory cycle may have on male-female associations, female partner preferences, and copulation rates

(Muller & Wrangham 2009; Muller et al. 2011). Finally, while Bercovitch did consider the effect of male rank and rank trajectory on female resistance behavior, he did not test what qualities might be consistently preferred among females more generally. Thus, I aim to use the olive baboons at Gombe Stream National Park, Tanzania to replicate

Bercovitch’s study on the same species at Gilgil, Kenya. However, I also make important expansions on his findings. First, I more systematically consider whether male aggression drives “preferences.” Second, I use modern techniques to narrow the fertility window and examine differences in behavioral patterns in different phases. Finally, I incorporate multiple behavioral measures to assess female partner preferences, assess how preferences impacts male-female copulation rates as they vary across the estrous cycle, and systematically test which male characteristics, if any, drive female preferences.

There are several lines of evidence that helped me develop a more comprehensive list of potential male characteristics that might be favored by females. First, Packer

(1980) demonstrated that male olive baboons are more likely to provide care to a

“friend’s” infant if they were present in the group when the infant was conceived.

Bercovitch (1991a; 1995) also suggested that mating with preferred partners can serve as an impetus for friendship formation in olive baboons. Thus, I predict the most valuable

47 mating partners for female olive baboons are males who are likely to be present in the social group when offspring are born and who will be best able to protect females and their vulnerable newborn offspring from predation or potentially harassment from new immigrant males (Henzi et al. 2010). I used tenure within the group prior to copulation as a proxy, and also followed up after my study ended to see who remained after a few years. Tenure in the group may signal quality in two ways. It may be, as Bercovitch

(1991) suggests, that new males with shorter tenures are more likely to remain in the group for longer periods of time. Alternatively, males with long tenure may be more familiar to females so there is more time to assess his potential as a future friend and/or his competitive ability. In addition to tenure length, males stable in their dominance rank or those ascending the dominance hierarchy are less likely than unstable or descending males to leave the group by the time a female’s infant is born and thus would be the best protectors of a female’s vulnerable infant. Therefore, the “highest” quality male mates are not just those of current high rank but those who will remain high ranked, those gaining rank in the social hierarchy, younger males, and those who have greater potential to remain in the group. These males should be consistently “preferred” by females.

I assessed preferences based on female-centric association indices (AIs) (Stumpf

& Boesch 2005; Girard-Buttoz et al. 2014) with males during ovulatory days. This time period is when female preferences are expected to be “strongest” because they would have the highest impact on her fitness (Stumpf & Boesch 2005). Following Stumpf and

Boesch (2005, 2006), I defined preferences from active interactions, directed by the female, toward some males but not others during her fertile phase (see Chapter 2). I then

48 searched for male characteristics that might explain female preferences and I predicted that females would be most interested in young, high-ranking males, either stable or ascending the hierarchy, whose tenure in the group had been long. Second, I compared aggression rates both within and outside the ovulation window between preferred and eschewed males to determine if male behavior drove female “choices.” Third, I tested whether female preference for a male might positively impact his frequency of copulations, his frequency of ejaculations, his ability to form a consort, whether he contested the consort of another male, and whether his consort challenge failed. I predicted preference would be important during the fertile period but that females would copulate less discriminately outside the ovulatory window.

Results

What accounts for variation in male-female dyadic associations?

Of the 179 possible male-female dyads in all groups across the entire study period, 89 interacted frequently (i.e., approached or left one another at least ten times)

(Soltis et al. 1997) (Table 3.1) and, of these, 54 engaged in consortships. Females, on average, interacted frequently with 58.2% of males available in their respective groups

(range: 33.3-100%; Table 3.2) and consorted with 45.3% of available males (range:

22.22-83.33%; Table 3.2). Despite interacting/consorting with approximately half of all males in the group on average, Association Indices (AIs) were only high (as defined in

Chapter 2) for 26.0% of the males in the group (Table 3.2). A partial correlation, controlling for hours of observation time, revealed that neither rank nor age of the female

49 was related to proportion of males with whom she interacted, consorted, or preferred

(Table 3.3).

Using a male-centric analysis, I tested which male characteristics correlated with the likelihood of interacting and consorting with more females. I found a relationship with age; young males interacted (rs= -0.559; p=0.001) and consorted (rs=-0.562, p=0.001) with a higher proportion of females than older males (Table 3.4; Figure 3.1). No other characteristics were related (Table 3.4). Although not part of my original hypotheses, the relationship with age is not particularly surprising, and likely reflects an attempt on the part of males to maximize mating success early in their reproductive lives

(Bercovitch 1991). However, this should not be taken as representative of female preference for younger males. Instead, it is important to consider whether there are similar correlations between female mate preferences and copulatory success or if copulation success is simply the outcome of male competition.

What qualities are associated with preferred male partners?

Of the 89 dyads who interacted frequently, 40 included preferred partners. In other words, 55.0% of those frequently interacting dyads included non-preferred males, suggesting females are not driving all interactions. On average, females had 1.9 preferred partners (range: 1-4). Relatively few males were classified as neutral (n=11, compared to eschewed: n=38 and preferred: n=40); therefore, in the following analysis, I compare differences in male characteristics between “preferred” and “not preferred” (i.e., lumping both neutral and eschewed males) partners.

In accordance with my original hypotheses, preferred males were ascending the dominance hierarchy (Mann-Whitney U= 1.705, one-tailed: p=0.044, two-tailed: p=0.088) and there were no other male characteristics that explained the preferred and non-preferred status of males (including male age, tenure, or rank) (Table 3.5). Using a generalized linear mixed model (GLMM) allowed me to compare relative effects and to control for individual contribution of multiple data points to the data set (because I included both male and female identity nested within group identity as random factors;

Chapter 2). This confirmed the above results: in the best reduced model (based on lowest

AICc; see Chapter 2), none of the variables had a significant effect on female preference

(Table 3.6).

Each of the above results consider female partner preferences across individual females. However, a male could be classified as preferred for one female and not for another. If males are preferred because of certain social qualities, like rank and rank trajectory, there should be some consistencies in association patterns and partner preferences across females in each study group. Figure 3.2 represents differences in relative association indices (i.e., deviation from female average AI) for each male in the groups during each study period. Small sample size precluded statistical analysis for these data because five or fewer females were followed in each group. Nevertheless, some interesting patterns emerged and can help explain results presented above.

Following Stumpf and Boesch (2005), values above 0 in Figure 3.2 indicate the male’s dyadic AI scores were higher than others in the group and that he was therefore more consistently preferred by multiple females. There was at least one male in each

51 troop who was more consistently preferred than others. Interestingly, in both AC and BA groups during the 2012 study period these were the lowest ranking males in their respective groups (AC: male SYR; BA: male SIM), while in AC group and DC group during the 2012 study period and BA group during the 2014 study period, these were the highest ranking males (AC: male LRA; BA: male SFI; DC: male RUK). Both low- ranking males in 2012 remained in their groups and at least one had gained rank by the

2014 study period (male rank data were not consistently available for AC troop following the 2012 study period). Females in DC group in 2014 expressed some consistent preferences for males in the top 50% of the dominance hierarchy. All 3 males with median AIs above 0 (AIS, ALA, HYT) were present in the group in 2012, although AIS, the highest ranking male in 2014, was just beginning to integrate in the group and was not included in the 2012 DC group data set. Seven of the eight males who had relatively high Association Indices across multiple females over the course of the study remained in their groups for at least 2 years after they were determined to be preferred partners

(Collins, pers. comm.).

Does female partner preference influence consorts?

Of the 54 dyads who engaged in consorts, 29 (53.7%) included preferred partners and 25 (46.3%) included non-preferred partners. Thus, females were frequently in consort with non-preferred males and there was no significant difference between the proportion of consorts females participated in with preferred and non-preferred partners (Wilcoxon signed-rank test: T = 0.130, p=0.896, N=19).

For 24 of the 54 male-female dyads who consorted (44.4%), other males followed and contested the consort. Partner preferences did not significantly predict whether or not males contested consorts (Pearson chi-square: χ2=0.611, p=0.434). In a GLMM testing possible predictors of whether a male contested a consort with a given female, only male rank and rank trajectory (Table 3.7) predicted contest behavior, suggesting this behavior is driven by male competition rather than female preference for specific male partners.

Although partner preferences did not influence whether a female consorted with a male or whether a male contested a consort, preferences did significantly influence whether the male that contested the consort was able to successfully takeover a consort during that cycle (Pearson chi-square: χ2=9.603, p=0.002; Figure 3.3). Partner preference also remained the only predictor of a successful consort challenge in a best reduced

GLMM testing the relative effects of partner preference along with other factors (Table

3.8), with male and female identities included as random factors. Of the dyads in which a male contested consorts with the female and another partner, 54.2% went on to engage in a consort themselves after the male contested. Of these, 84.6% involved a female and a preferred male partner.

Do females copulate more frequently with preferred partners and are copulations more likely to be complete?

Within the 89 frequently interacting dyads, males and females engaged in 281 copulations. Females copulated with individual males at an average rate of 0.18

53 copulations/hr during ovulatory days (range: 0-1.18 copulations/hr) compared to 0.08 copulations/hr during days outside the ovulatory window (range: 0-0.34 copulations/hr).

Among females, who engaged in at least 10 copulations over the course of the study (N=14), none of the individual female characteristics correlated with her average copulation rates during any of the contexts (i.e., within or outside fertility window, when in consort, or following complete copulations; Table 3.9).

Comparing copulation rates to male characteristics, males with longer tenure in the group copulated at higher rates during consorts (rs=0.499; p=0.021). During ovulatory days, males ascending the dominance hierarchy copulated at higher rates (rs=0.366; one- tailed: p=0.036; two-tailed: p=0.072). However, during non-ovulatory days, higher- ranking males copulated at higher rates (rs=0.361; one-tailed: p=0.038; two-tailed: p=0.076). No other male characteristics correlated with copulation rates in any context

(Table 3.10).

Among those dyads who engaged in consorts (N= 54 of 89) copulation rates were higher with preferred than non-preferred partners (Mann Whitney U= 3.314, p<0.001).

Additionally, among those dyads who interacted frequently during ovulatory days (N=50 of 89 dyads), there were significantly more copulations with preferred partners (average of 69.9% of copulations occurred with preferred partners; range: 20-100%) compared to non-preferred partners (Mann Whitney U= 3.968, p<0.001). However, among those dyads that interacted significantly during non-ovulatory days (N=44 of 89 dyads), there was no difference between copulation rates with preferred partners and non-preferred partners (Mann Whitney U= 1.378, p=0.168)

There was no significant difference between the proportion of copulations that were completed with preferred versus non-preferred partners inside (Mann Whitney U= -

0.379, p=0.704) or outside the ovulatory window (Mann Whitney U=-0.431, p=0.684).

Does male aggression influence patterns of associations and copulations?

Of the 219 aggressive interactions males directed towards females, 193 (88.2%) were non-physical threats and chases, and only 26 (11.9%) involved direct, physical contact (e.g., biting, shoving, and shaking). Only two of these physical attacks culminated in the male copulating with the female, and only one of these was a complete copulation.

Across the dyads observed interacting significantly, females received an average of

0.0617 instances of aggression/hour (range: 0-0.536 instances of aggression/hour) from males.

Female age and rank did not correlate with rates of aggression received within or outside the ovulatory window (Table 3.11). There was a slight trend in which young males directed higher rates of aggression on ovulatory days (rs=-0.355, p=0.089), but none of the other male characteristics correlated with rates of aggression directed (Table

3.12).

On ovulatory days, females did not receive higher rates of aggression from preferred partners (Mann-Whitney U=1.114, Ndyads=50, p=0.265). Furthermore, across the entire cycle (using those dyads who interacted both during ovulatory days and days outside the ovulatory window), rates of aggression did not differ between preferred and non-preferred partners (Mann-Whitney U= 0.609, Ndyads=27, p=0.570). However, on

55 non-ovulatory days, females received higher rates of aggression from preferred partners

(Mann-Whitney U= 2.419, Ndyads=44, p=0.016). Likewise, overall rates of aggression did not correlate with copulation rates during ovulatory days (rs=0.033, Ndyads=50, p=0.818), but did correlate with copulation rates during days outside of the ovulatory window (rs =

0.330, Ndyads=44, p=0.029)

Given that both individual males and females are represented in multiple dyads, these dyads are not independent. Therefore, I also utilized GLMM (in which I could include female and male identity nested within group to control for multiple samples from same individual as well as observation hours as random effect), to assess the relative effect of all variables described above (male and female characteristics, female preferences, rates of male aggression, female ovulatory status, whether in consort or not, and whether copulations were complete). These results confirm many results from univariate analyses described above. In particular, the best models indicated that female preferences as well as male rank trajectory predicted copulation frequency during ovulatory days and consort days (Table 3.13). During consort days, rates of aggression received also predicted copulation rates. During days outside of the ovulatory window, female preferences as well as rates of aggression received predicted copulations (Table

3.13).

Discussion

In this chapter, I tested 1) whether females prefer to interact with some males but not others 2) whether male characteristics were associated significantly with female

56 preferences for those males 3) did preferred males have significantly more consorts, copulations, or engage in more ejaculatory copulations with females than non-preferred and 4) do patterns of access significantly differ between the ovulatory window and outside (van Schaik et al. 2000; van Schaik et al. 2004) and 5) are patterns of consort formation or contests and copulation frequencies all be explained by the aggressive actions of males.

Results indicate that some males are preferred partners. Only about 25% of males, across all females reached the upper quartile of associations during ovulation (i.e., my definition of female “preference”), despite the fact that about 50% of all males in the group interacted/consorted with females during these days. When I combine female patterns, some males are clearly more preferred than others by multiple females.

However, although there is slight relationship with rank trajectory (which also appears as a predictor of copulation success during the ovulatory window), no other characteristics measured correlate with “preference” as I have defined it.

So females prefer some males over others, but based on what? It may be that rank trajectory is, in fact, a significant predictor of female mate preference, and that females can assess this via male participation in contests (Kitchen et al. 2005; Crockford et al.

2007). This was a relatively short study, and I may not have captured the long-term rank trajectory patterns that, if recognizable to females, would be most valuable to increasing reproductive success (i.e., assessing rank in a way that would predict if he will actually be present in the group when infants are young and vulnerable). Because rank trajectory was an important predictor of copulatory success, particularly during ovulation, and

57 potentially an indicator of female preference, it is important to consider ways to better assess rank and rank trajectory in primates. The current method to assess rank is to rely on male-male interactions outside of the context of direct competition for mates, which can be infrequent (Neumann et al. 2011). Among closely related chacma baboons, rank trajectory is strongly correlated with fecal testosterone levels (Beehner et al. 2006). This is reflected vocally and behaviorally in signals that males may assess when engaging in aggressive displays and contests (Bergman et al. 2005; Kitchen et al. 2009b) and females may assess when choosing mates. In the future, I will assess whether fecal testosterone concentrations in male olive baboons reflect rank trajectory (as in chacma baboons). This could be a useful physiological proxy for assessing social status along with current rating methods in studies like mine to tease apart the interacting influences of male competition and female mate preferences.

If rank trajectory is a predictor of female preference or copulation success, in part, because these males are more likely to be dominant when infants are born, it follows that females should prefer all males who will be present in the group for extended tenures regardless of their current rank. This is especially true in baboons because males of all ranks and ages form friendships with females (Smuts 1985). In olive baboons, unlike in other savanna baboons, males and females remain friends for protracted periods

(potentially as long as six years: Smuts 1985); thus, it should especially benefit females in this species to attempt to bias paternity, or at least mate more frequently with males who have qualities indicative of their ability to remain in the group for extended periods of time. I used age and current tenure as this proxy, assuming that males who have longer

58 tenures will be more stable in their position in the group and may already have formed bonds with female residents (Smuts 1985). Although tenure did influence copulatory success during consorts, neither age nor current tenure significantly predicted female mate preferences. However, when I combined females, and considered males that were more commonly preferred across females in the group based on relative associations

(Stumpf & Boesch 2005), the majority (seven out of eight) were those who remained in the group for at least two years after they were determined to be preferred partners

(Collins, pers. comm.). Thus, there may be some indicator to females of a male’s likelihood of remaining in the group that I could not parse out here.

Additionally, females may prefer males who were previous friends and

“protectors” of themselves and their offspring. In her “mating effort” hypothesis, Smuts

(1985) suggested that rather than providing males direct benefits in the way of caring for offspring, male friendships should instead provide males with preferential mating access to females in the future. It was not possible to assess current male-female friendships in this study because (a) friendships are “relaxed” outside of periods when females are lactating (Smuts 1985) and thus would not have been apparent in behavioral data described here and (b) the majority of focal females were either nulliparous, and thus had no opportunities to form friendships previously, or they had offspring that were older (> 4 years), thus most prior friendships would have dissolved (Smuts 1985). Anecdotally, only one male-female dyad fit the criteria to be classified as friends and I frequently observed the male in this dyad associating with and grooming the female’s two-year old infant.

This male was also classified as a preferred partner of that female. In both chacma

(Moscovice et al. 2009) and yellow baboons (Buchan et al. 2003), DNA evidence suggests males bias care towards those offspring whom they have sired. Because it can be difficult to collect DNA evidence and because paternity might not reflect individual behavioral patterns, endocrinological evidence could supplement studies and determine when individual males have switched to parental effort/friendships rather than mating effort. Shur (2008) has shown that decreased testosterone production in baboons followed the birth of a female friend, mirroring testosterone profiles of pair bonded species like callitrichids and humans. Thus, in the future it will be important to expand this study by examining friendships before and after the time period studied here and conduct longitudinal studies to address genetic (e.g., Moscovice et al. 2010) and hormonal (Shur

2008) indicators of paternity and other effects of long male tenures, and male-female sexual and social bonds at Gombe.

Finally, it is also possible that in the present study I have either failed to measure the characteristics, or suite of characteristics, that females are honing in on, or that these features are not measurable between preferred and non-preferred partners because they are highly variable across females (like friendship history with a particular male). The data suggest the latter, given that many females “preferred” certain partners but there seemed to be no clear pattern (with the possible exception being those males who are present for extended periods of time beyond the mating context). In some groups, females seemed to prefer high-ranking males and in others it was the lowest-ranking males who were preferred. Thus, it may be that other characteristics related to individual dyadic compatibility predict female mate preferences, such as markers indicative of genetic

60 dissimilarity (novel and or more distantly related), especially those genes associated the major histocompatibility complex (MHC) (Jordan & Bruford 1998). For example, in mandrills (Mandrillus sphinx), MCH dissimilarity predicts the likelihood that a male sires offspring with a female, and this may occur via either prezygotic or postzygotic selection mechanisms (Setchell et al. 2010). Finally, it may be that female-female competition restricts access to specific males and further limits female preferences. I address this more in Chapter 5.

Does partner preference or other male qualities predict reproductive access to females?

Although females did not engage in more consorts with preferred partners, females copulated more frequently with preferred partners around the days of likely ovulation when they happened to be in consort. Females were also more likely to engage in consorts with preferred partners who contested consorts compared to non-preferred partners who contested consorts. This is the first study to demonstrate an interaction between female preferences for specific males and success of a consort challenge. Only female partner preference predicted whether males successfully took over a consort. This seems surprising given that both rank and rank trajectory predicted who contested consorts. However, it is aligned with a recent analysis of the function of contesting in olive baboons (Danish & Palombit 2014), in which it was demonstrated that follower males of all ranks engage in both opportunistic and abandoned consort takeovers, the two most common types of takeovers follower males engaged in during the course of my study (J. Walz, unpubl. data). Interestingly, following a consorting pair and

61 contesting/challenging the male consort partner regardless of age and rank are important components of male-male competition in East African but not South African savanna baboons (Danish & Palombit 2014). It may be that this “alternative male reproductive strategy” (Gross 1996) is not just a function of male-male competition, but that female choice also plays a role in this unique behavior.

In addition to copulating more with preferred partners, females also tended to copulate more with males ascending the dominance hierarchy around the time of likely conception and more with males who had long tenures during consorts. These data suggest that females have some ability to determine which males are successful, under the constraints of male competition, and that being preferred by a female may “pay-off” for males, especially those who might not otherwise successfully copulate (e.g., low- ranking males or those who have recently immigrated). In stark contrast, only rank had a weak relationship with copulation success outside the ovulatory window. Thus, current rank only has a slight positive effect on male copulatory success, and then only on days when it may have less significant fitness consequences for the female (Stumpf & Boesch

2005, 2006). This aligns with previous investigations suggesting current male rank may not be as tightly correlated with reproductive success as in other species (Packer 1979b,

Bercovitch 1987, Danish & Palombit 2014).

Even if high-ranking males, those gaining rank, or those who have been present in the group for extended periods are not what consistently defines what it is to be a

“preferred” partner, it could still benefit females to mate with these males more frequently as part of “mixed-mating” strategy (i.e., mating with multiple males but some

62 more frequently than others during specific periods). Preferentially mating with “up-and- comers” to the dominance hierarchy around ovulation could confirm paternity in these males and encourage friendship formation. These males are less likely than current alpha males to be usurped and leave the group by the time a female’s infant is born and thus would make excellent friends for females in the future (Henzi et al. 2010). Thus, although not necessarily “preferred” based on my criteria (Chapter 2), it would likely be advantageous for females to cooperate in copulations with these males both during ovulatory periods and during consorts. This could be part of a strategy that allows females to “recruit” as many decent quality friends as possible for the future (i.e., if a female cannot get her “preferred” partner, she should next accept a male ascending rank or one that has been in the group for a long time). Packer (1980) suggested that male- female friendships can form even if males mate with females outside the “fertile” period

(although he based this exclusively on assessments of female sexual swelling cycle, rather than endocrinological data). Thus, my data reconcile how female olive baboons may attempt to “set-up” potential friendships with multiple males during mating periods by mating with multiple males, but still attempt to bias paternity towards specific, preferred partners.

Does aggression drive female preference and copulatory success?

Similar to findings by Bercovitch (1995), rates of aggression received from males in my study were relatively low, did not predict overall copulation success during the ovulatory window, and, although violent physical attacks by males constituted 12% of all

63 male aggression directed at females, rarely resulted in immediate copulations (2 of 281 observed copulations). Direct coercion (i.e., agonism that results in copulations) does not seem to be influencing the fact that females copulate more with preferred partners than other males during the ovulatory window, and also suggests that females had some control over copulations during this time. Furthermore, unlike in chimpanzees (Muller et al. 2011), even when I look at all cycling days, preferred males, who copulate most frequently during ovulatory days, are still not more aggressive. Thus, my results, together with Bercovitch (1995), suggest direct coercion is not an important reproductive strategy for males in olive baboons.

This is in contrast to the patterns outside the ovulatory period, when aggression rates do seem to influence copulatory success and preferred partners are more aggressive than non-preferred partners. However, preferred partners are not more likely to successfully copulate with females during this time. Given that preferred males are particularly aggressive outside the ovulatory window and copulate at higher rates around ovulation suggests further complexity to the interaction between male aggression, female preference, and copulatory success. Outside the ovulatory window, mating with an aggressive male may be a cost-effective strategy for a female to avoid immediate harm with low risk of conception, a pattern that has been observed in long-tailed macaques

(Macaca fasciucularis) (Engelhardt et al. 2006). Because preferred males were more aggressive outside ovulation window, it is possible that a female’s preference for him later is simply a delayed reaction and therefore part of the male coercive strategy (Smuts

& Smuts 1993). Alternatively, instead of preference for him during fertile window being

64 a concession for his prior aggression, perhaps females are actually attracted to males who have been aggressive previously. In particular, male aggressiveness may be a signal to the female that the male will be a good friend in the future; in other words, he may have high aggression with everyone, including others who threaten her or her offspring. In

East African baboons, males do not only defend offspring, but they also defend females from harassment from other females (Ransom & Ransom 1971; Altmann 1980;

Lemasson et al. 2008). Additionally, infants are vulnerable to chimpanzee hunting at

Gombe (pers obs., see also: Wrangham & Riss 1990) and this may be an added selective pressure driving female preference for both indicators of their likelihood to intervene

(based on aggression) or of their likelihood to be present in the group for extended periods of time.

Finally, it may be that higher rates of aggression reflect incidental aggression received as a result of associating more frequently; however, given that females did not receive more aggression from preferred partners around the ovulatory period, this seems less likely. Nevertheless, I plan to test this and further tease apart the relationship between male aggression, female partner preference, and copulatory success in the future using nearest neighbor data that I collected as well as rates of other social interaction data across the cycle.

In sum, my data support predictions postulated, but not exemplified in

Bercovitch’s (1995) study on olive baboons at Gilgil. Although females received aggression from males during days outside the ovulatory window, this behavior did not persist around ovulation and females copulated more frequently with their preferred

65 partners around the time when conception was most likely. Females also consorted with preferred males more frequently after these males challenged their current consort partner. Finally, my data demonstrate that, along with mating with preferred partners, females mate frequently with males rising in rank and males likely to be present in the group in the future, which could set the stage for multiple males to be available as friends of the female and her infant. While male competition is a primary driver of male-female dyadic and copulatory interactions in olive baboons, female choice has significant influence on the success of these interactions.

TABLES

Table 3.1: Copulatory and aggression rates while females were ovulating, non- ovulatory, and in consorts for 89 female-male dyads.

Nonov. Ov. Total Nonov. Ov. Con. Pref. Cop. Cop. Agg. Agg. Agg. Agg. Dyad (F-M) Year Troop Partner Rate Rate Rate Rate Rate Rate ULY-AMX 2012 AC no 0.00 0.24 0.04 0.00 0.49 0.04 UNK-AMX 2012 AC no 0.00 0.09 0.02 0.00 0.05 0.00 YLT-AMX 2012 AC no 0.11 0.13 0.00 0.00 0.00 0.00 UBG-AVZ 2012 AC yes 0.42 0.37 0.54 0.14 1.24 0.67 ULY-AVZ 2012 AC no 0.19 0.24 0.00 0.00 0.00 0.00 UNK-AVZ 2012 AC no 0.03 0.00 0.07 0.03 0.14 0.03 YLT-AVZ 2012 AC no 0.34 0.00 0.12 0.00 0.39 0.15 ULY-GBS 2012 AC yes 0.14 0.24 0.09 0.02 0.73 0.16 UNK-GBS 2012 AC no 0.03 0.09 0.10 0.08 0.14 0.19 UVA-GBS 2012 AC no 0.04 0.00 0.03 0.00 0.05 0.10 UNK-LRA 2012 AC no 0.15 0.09 0.10 0.15 0.00 0.16 UVA-LRA 2012 AC yes 0.04 0.49 0.10 0.09 0.12 0.29 ULY-SHT 2012 AC no 0.02 0.00 0.02 0.00 0.24 0.04 YLT-SHT 2012 AC no 0.00 0.00 0.00 0.00 0.00 0.00 UBG-SYR 2012 AC yes 0.00 0.25 0.54 0.63 0.37 0.67 ULY-SYR 2012 AC no 0.10 0.00 0.13 0.02 1.21 0.20 UNK-SYR 2012 AC yes 0.00 0.32 0.07 0.03 0.14 0.00 YLT-SYR 2012 AC yes 0.00 1.03 0.24 0.23 0.26 0.37 WDF-AGT 2012 BA no 0.00 0.00 0.00 0.00 0.00 0.00 WLD-AGT 2012 BA no 0.00 0.88 0.04 0.00 0.15 0.00 WTW-AGT 2012 BA yes 0.00 0.12 0.00 0.00 0.00 0.00 WDF-AST 2012 BA yes 0.08 0.16 0.08 0.03 0.24 0.13 WTW-RED 2012 BA yes 0.03 0.23 0.00 0.00 0.00 0.00 WDF-SFI 2012 BA no 0.01 0.00 0.02 0.00 0.08 0.00 WTW-SFI 2012 BA yes 0.13 0.23 0.07 0.06 0.12 0.24 WDF-SIM 2012 BA yes 0.06 0.20 0.02 0.00 0.08 0.00 WLD-SIM 2012 BA yes 0.10 1.18 0.00 0.00 0.00 0.00 WTW-SIM 2012 BA no 0.10 0.00 0.00 0.00 0.00 0.00 AKA-AGT 2014 BA yes 0.02 0.00 0.04 0.02 0.18 0.08 WGR-AGT 2014 BA no 0.00 0.00 0.00 0.00 0.00 0.00 AJA-AST 2014 BA no 0.00 0.00 0.05 0.00 0.35 0.17 Continued

Table 3.1 Continued Nonov. Ov. Total Nonov. Ov. Con. Pref. Cop. Cop. Agg. Agg. Agg. Agg. Dyad (F-M) Year Troop Partner Rate Rate Rate Rate Rate Rate WGR-AST 2014 BA no 0.00 0.00 0.06 0.00 0.20 0.00 AJA-SFI 2014 BA yes 0.08 0.17 0.02 0.00 0.17 0.17 AKA-SFI 2014 BA no 0.00 0.00 0.00 0.00 0.00 0.00 AJA-SIM 2014 BA no 0.00 0.00 0.00 0.00 0.00 0.00 AKA-SIM 2014 BA no 0.00 0.00 0.00 0.00 0.00 0.00 WGR-SIM 2014 BA yes 0.00 1.39 0.06 0.09 0.00 0.20 AJA-SUR 2014 BA no 0.00 0.00 0.00 0.00 0.00 0.00 AKA-SUR 2014 BA no 0.00 0.00 0.02 0.00 0.18 0.08 WGR-SUR 2014 BA yes 0.00 0.20 0.06 0.00 0.20 0.20 AKA-SUZ 2014 BA no 0.00 0.00 0.00 0.00 0.00 0.00 HON-ACH 2012 DC yes 0.14 0.29 0.20 0.09 0.67 0.16 HRF-ACH 2012 DC no 0.00 0.00 0.02 0.00 0.05 0.00 HRS-ACH 2012 DC yes 0.00 0.28 0.11 0.08 0.19 0.14 HRT-ACH 2012 DC yes 0.00 0.00 0.15 0.18 0.00 0.00 SAS-ACH 2012 DC no 0.00 0.00 0.00 0.00 0.00 0.00 HON-ALA 2012 DC no 0.02 0.00 0.06 0.00 0.29 0.00 HRF-ALA 2012 DC no 0.00 0.05 0.04 0.03 0.05 0.06 HRS-ALA 2012 DC no 0.14 0.09 0.07 0.00 0.28 0.07 HRT-ALA 2012 DC no 0.09 0.00 0.05 0.00 0.36 0.00 SAS-ALA 2012 DC yes 0.00 0.00 0.00 0.00 0.00 0.00 HRT-AMA 2012 DC yes 0.00 0.18 0.00 0.00 0.00 0.00 SAS-BU 2012 DC yes 0.00 0.12 0.03 0.05 0.00 0.17 HON-HYT 2012 DC yes 0.11 0.57 0.07 0.02 0.29 0.08 HRF-HYT 2012 DC no 0.00 0.15 0.08 0.09 0.05 0.18 HRT-HYT 2012 DC no 0.12 0.00 0.03 0.00 0.18 0.06 HON-RUK 2012 DC no 0.21 0.00 0.02 0.00 0.10 0.04 HRF-RUK 2012 DC yes 0.06 0.20 0.12 0.06 0.20 0.30 HRT-RUK 2012 DC no 0.00 0.00 0.00 0.00 0.00 0.00 SAS-RUK 2012 DC yes 0.05 0.23 0.17 0.05 0.46 0.50 HRF-UBO 2012 DC yes 0.00 0.20 0.04 0.03 0.05 0.06 HRS-UBO 2012 DC no 0.11 0.19 0.20 0.11 0.47 0.29 SAS-UBO 2012 DC no 0.00 0.00 0.00 0.00 0.00 0.00 HRS-A2 2014 DC yes 0.05 0.85 0.35 0.48 0.11 0.68 SAS-A2 2014 DC no 0.04 0.00 0.00 0.00 0.00 0.00 SER-A2 2014 DC no 0.00 0.00 0.03 0.00 0.12 0.00 Continued

Table 3.1 Continued Nonov. Ov. Total Nonov. Ov. Con. Pref. Cop. Cop. Agg. Agg. Agg. Agg. Dyad (F-M) Year Troop Partner Rate Rate Rate Rate Rate Rate SPR-A2 2014 DC yes 0.07 0.17 0.13 0.05 0.43 0.19 HRS-AIS 2014 DC yes 0.00 0.21 0.04 0.00 0.11 0.07 SAS-AIS 2014 DC yes 0.08 0.25 0.00 0.00 0.00 0.00 SER-AIS 2014 DC yes 0.00 0.24 0.06 0.04 0.12 0.07 HAU-ALA 2014 DC yes 0.01 0.35 0.00 0.00 0.00 0.00 SER-ALA 2014 DC yes 0.08 0.24 0.06 0.04 0.12 0.07 SPR-ALA 2014 DC yes 0.19 0.34 0.04 0.02 0.09 0.06 HAU-BEAR 2014 DC no 0.06 0.00 0.00 0.00 0.00 0.00 SAS-BEAR 2014 DC no 0.00 0.00 0.03 0.04 0.00 0.00 SPR-BEAR 2014 DC yes 0.00 0.00 0.02 0.00 0.09 0.00 HAU-CAN 2014 DC no 0.01 0.00 0.08 0.00 1.06 0.24 HRS-CAN 2014 DC no 0.05 0.00 0.04 0.00 0.11 0.07 SAS-CAN 2014 DC no 0.00 0.00 0.03 0.04 0.00 0.07 SER-CAN 2014 DC yes 0.08 0.24 0.03 0.00 0.12 0.00 SPR-CAN 2014 DC yes 0.12 0.26 0.20 0.02 0.86 0.22 HAU-HYT 2014 DC yes 0.00 0.35 0.00 0.00 0.00 0.00 HRS-HYT 2014 DC no 0.11 0.00 0.04 0.05 0.00 0.07 HRS-RUK 2014 DC no 0.00 0.00 0.00 0.00 0.00 0.00 SAS-RUK 2014 DC no 0.00 0.13 0.00 0.00 0.00 0.00 SPR-RUK 2014 DC no 0.02 0.09 0.04 0.00 0.17 0.00 SAS-SUP 2014 DC yes 0.00 0.25 0.06 0.08 0.00 0.15 SPR-SUP 2014 DC no 0.00 0.00 0.02 0.02 0.00 0.03

Table 3.2: Females and the percent of males with whom they interacted frequently, consorted, and “preferred.”

% Males % Males % Males Female Troop Interact Consort Prefer AJA BA 66.67 50.00 16.67 AKA BA 100.00 66.67 33.33 HAU DC 44.44 22.22 22.22 HON DC 44.44 33.33 22.22 HRF DC 55.56 44.44 22.22 HRS DC 44.44 38.89 16.67 HRT DC 55.56 33.33 22.22 SAS AC 61.11 38.89 27.78 SER DC 44.44 44.44 33.33 SPR DC 66.67 55.56 44.44 UBG AC 33.33 33.33 33.33 ULY AC 83.33 83.33 16.67 UNK AC 83.33 83.33 16.67 UVA AC 33.33 33.33 16.67 WDF BA 66.67 33.33 33.33 WGR BA 66.67 33.33 33.33 WLD BA 33.33 33.33 16.67 WTW BA 66.67 50.00 50.00 YLT AC 66.67 50.00 16.67

Table 3.3: Results of partial correlation controlling for total hours observed and female characteristics and the percent of males with whom females interacted, consorted, and preferred.

Context Rank Age

Total Interaction rs -0.175 -0.012 p 0.488 0.964 N 19 19 Consort rs -0.174 -0.105 p 0.490 0.678 N 19 19 Preferred rs 0.187 -0.196 p 0.541 0.522 N 19 19

Table 3.4: Results of correlation between male characteristics and proportion of focal females with whom they interacted, consorted, and were classified as preferred.

Context Rank Rank Age Tenure Trajectory Total Interaction rs -0.046 0.116 -0.599 0.121 p 0.829 0.579 0.002 0.583 N 25 25 24 23 Consort rs 0.044 0.120 -0.541 0.040 p 0.835 0.569 0.006 0.856 N 25 25 24 23 Preferred rs -0.052 0.281 -0.563 0.088 p 0.805 0.173 0.004 0.688 N 25 25 24 23

Table 3.5: Results of Mann-Whitney U comparing characteristics of preferred and non-preferred partners across all 89 dyads who interacted frequently during the study.

Male U N P value Characteristic Rank 0.903 89 0.367 Rank Trajectory 1.705 89 0.088 Tenure -0.915 89 0.965 Age 0.965 89 0.360

Table 3.6: Reduced model testing influence of male characteristics on female partner preferences.

Response Variable Fixed Factor Coefficient SE t value P value Partner preferencea Age 0.097 0.076 1.268 0.209 Rank -0.399 0.737 -0.542 0.590 Rank Trajectory -0.147 0.100 -1.465 0.147 aReference category: Not preferred

Table 3.7: Generalized linear mixed models with AICc values that best explain whether a male contests consorts among olive baboon females.

Response Variables included β±SE t p AICc Variable Contestsa Male Rank 1.589±0.811 1.960 0.054 365.392 Rank Trajectory -0.213±0.105 -2.024 0.046

aReference category: Yes

Table 3.8: Generalized linear mixed models with AICc values that best explain whether a male goes on to consort with females whose consorts they contested.

Response Variables β±SE t p AICc Variable included Male Consort Partner Preference 3.120±1.156 2.699 0.013 106.136 Following Contesta

aReference category: Yes

Table 3.9: Results of correlation comparing female rank and age to copulation rates during ovulatory periods, non-ovulatory periods, consorts, and complete copulations.

Context Rank Age

Ovulatory rs -0.194 0.246 p 0.507 0.396 N 14 14 Non-ovulatory rs -0.319 0.022 p 0.266 0.940 N 14 14 Consort rs -0.095 -0.009 p 0.748 0.976 N 14 14 Complete rs -0.337 0.167 p 0.239 0.568 N 14 14

Table 3.10: Results of correlation comparing male characteristics to copulation rates during ovulatory periods, non-ovulatory periods, consorts, and complete copulations.

Context Rank Trajectory Age Tenure

Ovulatory rs -0.119 0.366 -0.068 0.245 p 0.571 0.072 0.750 0.261 N 25 25 24 23 Non-ovulatory rs 0.361 0.203 0.006 0.307 p 0.076 0.330 0.977 0.154 N 25 25 24 23 Consort rs 0.141 0.190 -0.097 0.499 p 0.532 0.396 0.676 0.021 N 22 22 21 21 Complete rs 0.177 0.289 -0.186 0.205 p 0.398 0.161 0.384 0.349 N 25 25 24 23

Table 3.11: Results of correlation comparing rates of aggression female received during ovulatory and non-ovulatory periods from males with female rank and age.

Context Rank Age

Ovulatory rs -0.358 -0.191 p 0.133 0.434 N 19 19 Non-ovulatory rs 0.305 -0.314 p 0.205 0.190 N 19 19

Context Rank Rank Trajectory Age Tenure

Ovulatory rs -0.163 -0.073 -0.355 -0.048 p 0.437 0.727 0.089 0.827 N 25 25 24 23 Non-ovulatory rs -0.048 -0.086 -0.326 -0.196 p 0.819 0.681 0.120 0.370 N 25 25 24 23

Table 3.13: Generalized linear mixed models with AICc values that best predict copulations among olive baboon females during different social and reproductive contexts.

Response Variable included β±SE t p AICc Variable Ovulatory days Rank Trajectory -0.47±0.16 3.06 0.004 682.41

Partner Preferred 1.49±0.51 2.90 0.006

Non-ovulatory Partner Preferred -2.22±0.35 6.35 p<0.001 366.66 days Aggression Rates -13.18±2.25 -5.85 p<0.001

Consorts Rank Trajectory -0.48±0.11 -4.19 0.001 625.67

Partner Preferred 2.11±0.26 8.19 <0.001

Aggression Rates -2.381±1.15 -2.07 0.043

FIGURES

Figure 3.1: Spearman correlation between male age and the proportion of focal females with whom males were observed interacting.

Figure 3.2: Boxplots showing variation in relative association indices (pooling all individual female data) across males in each study group. The three-letter codes represent individual males. Values above 0 indicate the male in question’s dyadic AIs were higher than others in the group and was more consistently preferred by females. Males are ranked from low to high, left to right, in their respective groups with regards to dominance rank. (a) AC group from the 2012 study period. (b) BA group from the 2012 study period. (c) BA group from the 2014 study period. (d) DC group from the 2012 study period. (e) DC group from the 2014 study period.

(a)

(b)

continued

Figure 3.2: Continued

(c)

(d)

(e)

Figure 3.3: Percent of times a consort was challenged by a preferred or non- preferred male and percent of times the challenge was successful (i.e., result in a consort takeover).

CHAPTER 4: DO FEMALE COPULATION CALLS ENCOURAGE MATE GUARDING FROM PREFERRED PARTNERS?

Introduction

In the previous chapter, I described the positive influence of female partner preferences on copulation rates and on consort takeovers. I also demonstrated a possible delayed relationship between female partner preference and male aggression. Thus, these associations may reflect male coercion rather than female choice. To further distinguish between these two hypotheses, it is important to expand this investigation to also consider potential “post-copulatory” behavioral signals of female mate preferences, like copulation calls and “darting” (O’Connell & Cowlishaw 1995), which could influence the likelihood of future copulations.

Copulation calls are rhythmic vocalizations expressed during mating among many catarrhine primates with multi-male, multi-female social organization (Dixson 1998; van

Schaik et al. 1999). In baboons and macaques, copulation calls are relatively low- frequency grunts, while in other species, like chimpanzees (Pan troglodytes), bonobos

(Pan paniscus), and talapoins (Miopithecus talapoin) they are high-frequency screams

(reviewed by Maestripieri & Roney 2005). Likelihood of call production also varies significantly across species, with females in some species producing calls as often as following 98.8% of copulations (i.e., pigtail macaques, Macaca nemestrina: Gonzoules et al. 1998) and others producing calls following less than 10.0% of copulations (e.g.,

Japanese macaques, Macaca fuscata: Oda & Masataka 1992; Tonkean macaques,

Macaca tonkeana: Aujard et al. 1998).

At least 15 different hypotheses have been proposed to explain the function of copulation calls, or lack thereof, and, although highly variable, most share a common framework: in a variety of species, copulation calls increase reproductive success among females and thus are sexually selected (Maestripieri & Roney 2005; Pradhan et al. 2006;

Townsend et al. 2011). Most of the hypotheses with empirical support can be further combined into either male competition/paternity confusion or female choice/mate- guarding hypotheses. One advantage of these adaptive hypotheses is that they predict one of the most fundamental characteristics of these calls – call timing relative to copulation

(Pradhan et al. 2006). Both hypotheses emphasize the adaptive benefits of calls as occurring through post-copulatory sexual selection processes (Birkhead & Pizzari 2002;

Pradhan et al. 2006). The data, at least across many primate species (Pradhan et al. 2006), support this prediction; females generally produce calls during the final stages of a mounting event and commonly continue vocalizing through the period of time immediately following copulations when they are moving away from a male (Dixson

1998). Most other hypotheses put forward to explain calling deal with proximate/causal explanation such as the idea that calls are just byproducts of intercourse (Hamilton &

Arrowood 1978) or stimulators of ovulation (Cheng 1992). These predict that calls should occur prior to copulation or in conjunction with male ejaculation, which does not account for the above observations in primates. However, proximate and ultimate/adaptive explanations are not mutually exclusive (Tinbergen 1963). Calls may

81 function in an adaptive fashion, while still being produced by mechanisms such as orgasm (reviewed in Puts & Dawood 2006). In fact, proximate explanations like those that suggest calls are brought on by hormone and muscle interactions nearer to ovulation

(Puts & Dawood 2006) dovetail nicely with functional/adaptive explanations where male listener’s (be they the copulating partner or other males in the group) use this information because it also advertises receptivity.

I focus here on these two adaptive explanations, male competition/paternity confusion and female choice. These hypotheses differ essentially in their explanations of how calls benefit females. According to the female choice hypothesis, calls are directed at the current partner to encourage certain partners to mate-guard or continue copulating

(Todt et al. 1995), thereby reducing the likelihood of copulating with other partners and increasing paternity estimates in males preferred by the female (Maestripieri & Roney

2005). Alternatively, females may use calls to advertise receptivity to a wider audience and thus incite male competition either directly, where the winner of a contest gains mating access to a female and the female consequently mates with “higher-quality” partners (Cox and LeBoeuf 1977; O’Connell & Cowlishaw 1994; Nikitopoulos et al.

2004), or indirectly, where calls encourage multiple males to mate with the female and provokes scramble competition at the level of the males’ sperm (O’Connell & Cowlishaw

1994). Encouraging multiple males to mate may additionally confuse paternity and thereby benefit females in reducing the likelihood that they will be coerced in the future through infanticide (O’Connell & Cowlishaw 1994).

These two hypotheses have some overlapping and some unique predictions (see

Table 4.1). Only recently have researchers begun to test between these hypotheses producing equivocal results. Data on Barbary (Macaca sylvanus) and long-tailed macaques (Macaca fascicularis) provide an example of how data from the same species might align with predictions of each hypothesis. Following copulation calling in both macaque species, male partners are more likely to stay close to a female and mate with her again, females are unlikely to solicit interactions with other males, and other males are unlikely to attempt to disrupt the consort (Todt et al. 1995; Nikitopoulos et al. 2004).

All these results support the female choice hypothesis. Conversely, the time between copulations is shorter when female Barbary macaques produce calls than when they do not, but not necessarily with specific partners, indicating calls increase copulation frequency irrespective of partner identity (Semple 1998). This result supports the competition/paternity confusion hypothesis. Furthermore, copulation calls among both

Barbary (Pfefferle et al. 2008) and long-tailed macaques (Engelhardt et al. 2012) signal the occurrence of ejaculation but not whether the female is ovulating and thus conceptive, suggesting calls communicate to non-mating partners. Avoiding advertising ovulation should limit opportunities for individual males to monopolize mating with females and increase opportunities to either confuse paternity or incite sperm competition.

One alternative explanation (Pradhan et al. 2006) for this ambiguity is that both strategies are at play and that females alter behavior based on whether or not they are nearing ovulation. This model assumes copulation calls, like sexual swelling sizes, are graded signals that both confuse and confirm paternity (Nunn 1999) and depend on the

83 make-up of the audience and, potentially, the characteristics of her partner such as rank.

Additionally, other female characteristics, beyond just her stage in the ovulatory cycle, might affect calls. For example, female chimpanzees vary their call production based on their age and parity status; young nulliparous females are less discriminate in their call production, which Fallon and colleagues (2016) suggest is a function of a reduced threat of coercion via infanticide. They speculate that young females with no small offspring are able to more “freely” advertise their sexual receptivity, instigate sperm competition, and increase their likelihood of conceiving.

The best way to examine competing adaptive hypotheses about copulation call production would be to compare differences in patterns among closely-related species that differ in subtle aspects of their social/mating systems. The three savanna baboon species are ideally suited for such a comparison. They live in multimale, multifemale social groups in which females mate polygynandrously (Melnick & Pearl 1986). Sperm competition is suggested to be significant in these species given that males express high ejaculation rates and testicular size is relatively large compared to body mass (Bercovitch

1989; Harcourt 1997). Yet, mate guarding is common in all three species and may function to limit opportunities for other males to copulate with females around the time of likely conception (Birkhead & Pizzari 2002). Males may coerce females into mating, either through directed aggression or indirect infanticide, and female savanna baboons may be limited in their ability to have direct choice in mating partners (Maestripieri &

Roney 2005). Given the influence of sexual dimorphism and male-female aggression in baboons, a female’s only opportunity for subtle influence on mating partners is via

84 copulation calls and other peri-copulatory behaviors (Maestripieri and Roney 2005). The sexual conflicts that arise should lead to the evolution of mechanisms following copulation that could increase female reproductive success, like either concentrating paternity via maintaining consort relationships or confusing paternity/inciting sperm competition through multiple mating.

There is some evidence indicating copulation calls in baboons evolved through post-copulatory mechanisms of sexual selection. Compared to studies on other primates, like macaques (Engelhardt et al. 2012) and chimpanzees (Townsend et al. 2008, 2011), recent studies from baboons indicate that copulation calls do vary according to female ovulatory cycles (chacma, Papio ursinus: O’Connell & Cowlishaw 1994; olive, Papio anubis: Rigaill et al. 2013). Among female chacma baboons, calls are longer around maximal swelling tumescence (O’Connell & Cowlishaw 1994), in olive baboons females call more frequently around ovulation (Rigaill et al. 2013), and in yellow baboons (Papio cynocephalus) the number of units per call, call duration, and rate of unit production all vary with sexual swelling size (Semple et al. 2002). While this could be taken as evidence of attempts to advertise the likelihood of conception with current copulatory partners consistent with expectations of the female-choice hypothesis, it could also be evidence that calls are encouraging other males to copulate to increase their paternity estimates or allow for sperm competition. Thus, other lines of evidence must be explored to tease apart the possible function of copulation calls in baboons (Maestripieri & Roney

2005).

Although largely similar in their coarse social systems, there are subtle differences among the three savanna baboons. Chacma baboons are considerably more coercive to females than in the other two species (reviewed in Kitchen et al., 2009a); for example, infanticide is the leading cause of infant mortality in chacmas, but relatively rare in the yellow and olive baboons (Palombit et al. 2000; Henzi & Barrett 2003). As predicted, there is evidence that female chacma copulation calls function to encourage male competition/paternity confusion in chacma baboons (O’Connell & Cowlishaw

1994). Female chacma baboons calls are very loud (audible at >100m: O’Connell &

Cowlishaw 1994), exaggerated in length (see Figure 4.1) and produced indiscriminately following 83% (O’Connell & Cowlishaw 1994) to 97% (Saayman 1970) of all copulations. In chacma baboons, male rank is tightly correlated with their ability to monopolize females in consorts. Therefore, females may only be able to discourage infanticide from lower-ranking or recent immigrant males if they can mate with multiple males following a consort with a higher-ranking male or if they encourage “sneak” copulations (Crockford et al. 2007), which occur frequently in the population.

Experimental evidence by Crockford and colleagues (2007) demonstrates that non- consorting males monitor consorting pairs even from great distances (up to 100m) and take advantage of opportunities for sneaky copulations (i.e., when consorting pairs sound as though they are temporarily separated). Mechanisms that encourage copulations with multiple male partners, such as the production of copulation calls or “darting” for great distances (O’Connell & Cowlishaw 1995; see Discussion), may be particularly important in chacma baboons.

On the other hand, reports from yellow baboons suggest they encode information on the rank of their copulating partner, producing longer calls with more units after mating with males who are ranked higher in dominance hierarchies (Semple et al. 2002).

Semple (2001) has also demonstrated through playback experiments that male yellow baboons respond to calls of consort partners, but show no response to calls of non- consorting females suggesting calls increase interest in the current partner. Although encoding information on male rank may be in agreement with both the male competition hypothesis and female choice hypothesis, the lack of interest in “available” females goes against the predictions of the male competition hypothesis. Semple (2001) concludes that yellow baboon copulation calls function in female mate choice.

Female olive baboons are less promiscuous and mate with fewer males on average per cycle than in either chacma or yellow baboons (reviewed by Pradhan et al. 2006) and female olive baboons produce copulation calls significantly less often (ranging from

19.0-62.0%: Ransom 1981; Bercovitch 1985) than in either of the other species (yellow baboons: 80.0- 96.9% of the copulations: Collins 1981; Semple 1998: chacma: above).

On the other hand, olive baboons are more similar to yellow baboons with respect to several characteristics of their social structures and reproductive strategies. In particular, male monopolization ability is not as tightly correlated with male rank in either species as it is in chacma baboons (Strum 1982; Smuts 1985; Bulger 1993; Alberts et al. 2003).

Additionally, whereas groups of male rivals will follow a consorting pair in yellow and olive baboons (Danish & Palombit 2014), a consorting pair is given a wide berth in chacma baboons. Furthermore, the copulation calls of both yellow and olive baboons are

87 much quieter (Semple 2001; Kitchen pers. comm) and shorter than in chacmas (Figure

4.1). Thus, we might expect that female copulation calls serve similar functions in both yellows and olives.

Lower calling rates and the quieter, shorter form of the copulation calls in olive baboons suggest they are less likely to function in encouraging male competition/paternity confusion, but may function to encourage mate-guarding and serve as a post-copulatory mechanism of female choice. Nevertheless, no studies on copulation calls in olive baboons have incorporated a large sample size from a wild population to look at multiple features of individual calls (frequency of call production as well as call duration, units/call, etc.) and assess whether calls vary according to female partner preferences, timing with respect to ovulation, and copulatory context (as in Table 4.1).

If females call frequently following ejaculatory copulations, during days surrounding ovulation and following complete copulations (i.e., with ejaculation), this will demonstrate that these calls play a role in post-copulatory sexual selection. If calls encourage competition/paternity confusion, females should call most often during copulations when “listener” males are present and there should be a positive relationship between calling frequency and number of partners. If calls encourage mate-guarding from

“preferred” partners, I predict females should call most frequently during copulations with these males or with males who are otherwise high ranking, during copulations in which other males are not in close proximity to the copulating pair, and there should be a negative relationship between calling frequency and number of partners. If females call more frequently in the absence of “listener” males then it is likely that there is

88 information encoded in the call that is communicated to the current mating partner.

Alternatively, if females are employing different strategies with different males, females copulating with partners who are not preferred should call more frequently when

“listener” males are present (within close proximity). Additionally, I predict that the form of the calls (e.g., in terms of duration, number of “units,” and other acoustic features:

Semple et al. 2002) vary according to the female’s preference for the male, the presence of listeners, and based on stage in her ovulatory cycle. For example, exaggerations should occur when no listeners are present if the features function to encourage partner males to continue mate-guarding or when listeners are present if these features function to increase competition. Finally, if calls encourage mate-guarding, I predict that females who call often will engage in longer consorts.

In addition to copulation calls, all female savanna baboons “dart” following copulation, running away quickly for up to several meters before the male has dismounted (Collins 1981; Smuts 1985). O’Connell & Cowlishaw (1994, 1995) proposed that likelihood or distance of darting might function in female mate choice by encouraging the copulating partner to follow the female and consequently form a consortship or darting might function in competition/paternity confusion by assisting other males in finding the copulating female. In their study of chacma baboons, however, they were unable to reject either hypothesis as they found that females darted farther when close to ovulation but there was no effect of ejaculation and only a minor effect of adult versus juvenile partners that might have been confounded by ovulation state

(O’Connell & Cowlishaw 1995). Unlike their study, I will test for the function of darting

89 using only adult males with known rank, rank trajectory, age, tenure length, and preference data.

Results

Calling frequencies and comparisons to other studies

Adult female subjects called on average during 50.1% (range 0-92.86%) of the

281 copulations with adult males in which it was clear whether the female called and whether or not the copulation was complete (Table 4.2). This is similar to recent reports from other sites, including CNRS Primatology Station, France (47.4%: Rigaill et al.

2013) and Gilgil, Kenya (62.0 % during copulations with adult males: Bercovitch, 1985), but differs from the19.0% reported in Ransom’s (1981) comprehensive study on olive baboons at Gombe. Ransom conducted his fieldwork from 1967-1970, during the early stages of habituation at Gombe. The difficulty of maintaining close following distances combined with the quiet nature of copulation calls in olive baboons may explain the low calling rates in his original study.

Is there an influence of male and female characteristics on calling frequencies?

Among females whose calling was discernable and who copulated at least 10 times over the course of the study, neither female rank (rs = 0.248, N=13, p=0.414) nor female age (rs = 0.330, N=13, p=0.029) correlated with calling frequencies. Calling also did not vary between nulliparous and parous females (Mann-Whitney: U=0.286, p =

0.836).

None of the male characteristics I considered correlated with female calling frequency, including those variables that positively loaded on the male sociality measure

(e.g., relative rank and rank trajectory; Chapter 2) (rs = -0.175, N=24, p=0.414), or those variables that loaded on the experience measure (e.g., age and tenure in the group) for those males whom age and tenure were known (rs = -0.243, N=21, p=0.289). Because no variables demonstrated a relationship, I do not consider them further with regard to calling frequency.

Does the copulatory context explain variation in female calling behavior?

Whether or not the male ejaculated significantly influenced female calling

(Wilcoxon signed-rank test: T = -2.499, p=0.012, N=11; Figure 4.2). On average, females called during 63.9% (range 0-100%) of copulations in which a male ejaculated, but only called during 19.5% (0-100%) of copulations in which a male did not ejaculate. Only two females (WLD and WTW) called less often during complete copulations than during incomplete copulations and one female never called during any copulations (YLT; Figure

4.2). Because female calling starts following male ejaculation in olive baboon (unlike in

Barbary macaques: Pfefferle et al. 2008), we considered this relationship directional.

Females called on average following 36.1% of copulations outside of consorts

(range: 0-75%) and following 54.8% of copulations during consorts (0-93%). Within females, the difference in calling during consorts compared to outside of consorts did not reach significance, but this likely reflects a small sample size problem (T =1.599,

91 p=0.110, N=9). Copulations outside of consorts were relatively rare and four females never copulated outside of consorts, reducing my sample size to nine females.

Within consorts, females called significantly more often when a male was not contesting the consort than during contested consorts (T=-2.100, p=0.036, N=10; Figure

4.3). Females called on average following 59.0% of copulations during uncontested consorts (range: 0-100%) and following 30.1% of copulations during contested consorts

(range: 0-75%). Only one female called more often during contested consorts (HRS) and two females never called during any consorts (WLD and WTW; Figure 4.3).

Calling more following complete copulations and in the absence of a contesting male supports predictions of hypotheses generally suggesting calls play a role in post- copulatory female choice. In fact, given the low-amplitude nature of copulation calls in this species, if calls function to encourage competition, particularly sperm competition, it is likely that competitor males would have to be in relatively close proximity of the copulation to perceive the calls. Based on this assumption, I broadened my test to further considered whether females, following completed copulations, but regardless of the mating context (consortship or unguarded), were more likely to call in the presence (i.e., within 10m) of any male other than the copulating partner (i.e., not restricted to contesting males). Females tended to call more often in the absence of any listener males

(T=-1.859, p=0.063, N=10; Figure 4.5).

There was no significant effect of female ovulatory status, assessed through both fecal estradiol (T=0.235, p=0.814, N=13) and sexual swelling size (T=0.153, p=0.878,

N=11) or mate preference (T=-0.889, p=0.374, N=12) on female calling.

To examine the relative effect of all these predictor variables and to control for multiple samples from the same individuals, I ran a Generalized Linear Mixed Model

(GLMM). I tested whether the fixed factors of female partner preference, female ovulatory status, copulatory context (i.e., consort or not), presence of contesting male, and occurrence of ejaculation predicted calling frequency, with female and male partner identity nested inside group identity as random factors. In the best reduced model (i.e. with the lowest AICc values; Chapter 2), I found that both ejaculation and whether a contesting male were present predicted calling (Table 4.3).

What information is incorporated into individual calls?

To better assess the information incorporated in the calls themselves, I considered several temporal features of the calls, including the number of call units produced within a call, the overall call duration, unit rate (units/second), number of inhales, number of exhales, and exhalation rate (exhalation units/second). Based on sound quality, this was possible in 53 calls from 13 individual females (average number of recorded calls:

4.08/female SD: 2.87).

The continuous variables (call duration, unit rate, and exhalation rate) were tested for normality. Call duration was not normally distributed (skewness=2.169±0.327, kurtosis=6.428±0.644), but unit rate was normally distributed (skewness=0.360±0.327, kurtosis=-0.568±0.644) as well as exhale rate (skewness=0.248±0.327, kurtosis=-

0.348±0.644). Following log transformation, call duration was normally distributed (call duration: skewness=0.081±0.327, kurtosis=-0.688±0.644). Each of the call features were

93 tested for effects of male and female characteristics, including whether or not females were fertile. Most high quality calls were recorded following ejaculation, during consorts, and in the absence of a neighboring male. Therefore, these variables were not included in this analysis.

For each call feature, I considered whether characteristics of females and males correlated with features of female calls. None of the female characteristics I considered correlated with any of the call features (Table 4.4). Unit rate and male sociality correlated at a level that approached statistical significance (rs=0.473, N=17, p=0.055; Table 4.4).

Sociality is a composite index that is heavily weighted by male rank and rank trajectory

(see Chapter 2); therefore, males of high ranking males increasing or stable in rank had faster unit rate per call than other males (Figure 4.5). There seemed to be one outlier, who was a male with an abnormally high rank trajectory score and corresponding high sociality index that does not follow the trend (Figure 4.5). Removing this male from the analysis leads to a significant positive correlation between male sociality and unit rate

(rs=0.539, N=16, p=0.031). None of the other male identity characteristics I considered approached statistical significance. Thus, I include only male sociality in models described below.

Whether or not females were ovulating predicted the number of units produced per call (Table 4.6; Figure 4.6) as well as the number of exhales per call (Table 4.6;

Figure 4.7) with females producing calls with more units, particularly more exhales, during days surrounding ovulation. However, the apparent influence of female fertility on call unit production may have reflected an extreme outlier, an individual female whose

94 call included significantly more units than the mean number of units incorporated into the calls of other ovulating females (Figure 4.7).

Females produced shorter calls with eschewed males and longer calls with preferred and “neutral” males (Table 4.6; Figure 4.8). Although the number of units did not vary with partner preference (Table 4.6), there was trend in which females produced calls with faster exhale rates (Table 4.6; Figure 4.9) and faster unit rates generally (Table

4.6) following copulations with eschewed males. Although significant alone, male sociality versus unit rate was only a trend.

Do calls influence the number of partners, the likelihood of pair consorting in the future, or the length of their consorts?

Only 35 copulations took place outside of the context of a consort, and these were distributed across 20 male-female dyads who copulated on average 1.8 times (range: 1-

4/dyad). Unsurprisingly, based on patterns observed across all copulatory contexts, females generally called more frequently during complete copulations compared to incomplete copulations outside of consorts (Pearson Chi-Square: χ2=5.544, p=0.019).

During these unguarded contexts, there was no difference between female calling while pre-ovulatory, ovulatory, or post-ovulatory (χ2=0.113, p=0.945) or while copulating with preferred versus eschewed partners (χ2=0.793, p=0.373).

Of the 20 male-female dyads that engaged in copulations outside of the context of a consort, seven were then observed forming a consort later within that female’s menstrual cycle. However, of these seven dyads, only three engaged in copulations in

95 which the female produced a copulation call. Thus, due to small samples sizes, it is difficult to statistically assess whether calling during unguarded copulations increases the likelihood of engaging in consorts in the future and whether calling more frequently with preferred partners outside of consort encourages males to mate-guard in the future.

However, there is some evidence that calling following copulations while already in a consort is related to the duration the consort is maintained. Among the 41 male- female dyads who engaged in copulations during consorts, there was a positive correlation between female calling frequency and number of days spent in consort with that same male (rs = 0.321, p=0.040; Figure 4.10).

Finally, if female calling rates encourage multiple male partners, then females who call frequently should have more partners than females who call very little on average. In contrast, I found that females that called more frequently had fewer average male partners (step-wise regression, r2 = 0.50, F = 11.98, p = 0.005; Figure 4.11).

Although this is correlational and thus I cannot ascertain cause and effect, this result is more aligned with the female choice/encourage mate guarding hypothesis (Table 4.1).

Do male or female characteristics influence darting distance following copulations?

Females darted following 78% copulations and traveled an average of 4.3 meters

(range: 0-8.3 meters/copulation). They did not dart significantly different distances when copulating with preferred versus non-preferred partners (Wilcoxon signed-rank test: T = -

0.524, p=0.600, N=13). However, following complete copulations (i.e., with ejaculation), females darted significantly farther than during incomplete copulations (Wilcoxon

96 signed-rank test: T = -2.275, p = 0.023, N=12; Figure 4.12). Restricting analyses to just these complete copulations, the best reduced model revealed that none of the male characteristics (including whether the partner was preferred) significantly predicted dart distance; however, female age, parity, and ovulatory status did predict dart distance following complete copulations. Older, multiparous females traveled shorter distances following a copulation (Table 4.7; Figure 4.13) and females had shorter dart distances during copulations that occurred within the ovulatory window (Table 4.10; Figure 4.14).

Discussion

In this study, I analyzed both temporal and acoustic features of copulation calls in olive baboons to test hypothesis associated with whether calls vary with respect to copulatory context, qualities of male copulatory partners, and endocrinologically determined stages of female ovulatory cycles. My results show that ejaculation and audience presence best predict calling frequency in olive baboons and that females are more likely to call following “complete” (i.e., ejaculatory) copulations when competitor males are not present. Furthermore, females who call more frequently have fewer overall partners and remain in consorts for longer periods of time. Although my results do not indicate females call more frequently depending on their ovulatory status, I found some support that they encode information about ovulation in call features like the number of units, particularly exhales. Female olive baboons also seem to encode information about the status of their partners in their calls, producing calls with higher unit rates with high- ranking males who are also either maintaining rank or increasing in rank (i.e., males with high sociality scores; see Chapter 2). Finally, females encode information about their

97 preference for males, or lack thereof, by producing shorter calls with faster exhale and overall unit rates following copulations with eschewed partners.

Results from this study are the first to demonstrate that copulation calls in olive baboons are likely under selective pressures associated with post-copulatory mechanisms of sexual selection and may increase female likelihood of remaining in consorts with certain males (i.e., the female choice/encourage mate-guarding hypothesis; Table 4.1). As predicted, with the exception of calling rates – lower in olive baboons than both other savanna baboon species – olive baboon copulation calls seem to function more like they do in yellow baboons, as attempts to encourage consortship formation/continuation with the current partner (i.e., female mate choice: Semple 2001; Semple et al. 2002). In contrast, studies of chacma baboons suggest that their copulation calls function more in inciting male competition (O’Connell & Cowlishaw 1994; Crockford et al. 2007). Given the subtle but important differences in the social and mating systems of chacma versus olive and yellow baboons (see Introduction), particularly the high threat of infanticide that would make paternity confusion a great advantage to female chacma baboons, this is a plausible hypothesis that needs to be systematically tested in a comparative study.

Further differences between chacma and olive baboons can be seen when comparing their darting behavior. Female olive baboons who were older, multiparous, ovulating and following incomplete copulations were more likely to cover short dart distances following copulation than younger, primiparous/nulliparous,non-ovulating females following ejaculation. Based on hypotheses proposed by O’Connell and

Cowlishaw (1995), this might suggest that long darting distances function to incite

98 competition (by drawing attention to the copulating female) but that females only use this strategy (which would confuse paternity), when least likely to conceive or when less likely to have experience with male coercive tactics (see Fallon et al. 2016). O’Connell and Cowlishaw (1995) examined darting in chacma baboons and found a different result

–females darted farther distances when closest to ovulation regardless of ejaculation or of their partner or their own characteristics, which does not clearly align with either hypothesis (Table 4.1). Thus, my results suggest more systematic comparative studies are necessary and also demonstrate further complexity in these post-copulatory behaviors associated with female reproductive history.

Copulation calls in olive baboons play a role in post-copulatory sexual selection

There has been some debate about whether or not copulation calls by female catarrhines have a function or are simply brought on by anatomical mechanisms associated with orgasm (Maestripieri & Roney 2005). Additionally, in a variety of species there is no connection between calling frequency or call acoustic features and phases of the ovulatory cycle (Pfefferle et al. 2008; Townsend et al. 2011; Engelhardt et al. 2012). For copulation calls to be characterized as possible post-copulatory, sexually selected signals, Maestripieri and Roney (2005) suggest that females should produce calls more frequently during periods in which conception is most likely and following ejaculatory copulations. Results presented here demonstrate copulation calls in olive baboons are strongly associated with male ejaculation, and, although they do not vary in frequency with ovulation there is some evidence that they vary in form with ovulation.

Furthermore, in the only other study to assess changes in calling frequency according to likelihood of conception in olive baboons, Rigaill and colleagues (2013) demonstrated that female olive baboons produce more calls during the fertile phases of females estrous cycles compared to both pre-fertile and post-fertile phases. They incorporated photographic techniques to precisely measure changes in female swelling height, width, and labial width (Rigaill et al. 2013), which Higham and colleagues (2008) have validated as an accurate and precise method for assessing timing of ovulation.

Synthesizing results from this study with Rigaill and colleagues (2013) suggests that female olive baboons indicate their ovulatory status both in call form and frequency.

Similarly, I found that darting behavior varies with both ejaculation and phase of the cycle. This seems to suggest that darting and copulation calls can be included with sexual swelling size variation as a complex, graded signal of potential fertility following the predictions of Nunn (1999).

Teasing apart the adaptive hypotheses of copulation calls in olive baboons

Several lines of evidence from my study can help disentangle whether the female choice/mate-guarding or male competition/paternity confusion hypotheses fit copulation call production in olive baboons. Although female olive baboons call frequently following ejaculatory copulations, completeness of the copulation is not a “perfect” predictor of female calling in olive baboons compared to species like long-tailed macaques, who have been observed calling following 99.6% of ejaculatory copulations

(Engelhardt et al. 2012). Because of this, I assessed what might lead females to avoid

100 producing calls following complete copulations and demonstrated that females were less likely to call following these ejaculatory copulations if any other males were nearby (be they contesting or any nearby males). In fact, four of the ten females who called during ejaculatory copulations never called in the presence of a neighboring male. It is difficult to reconcile this observation with the male competition hypothesis for female copulation calls and thus it seems likely that copulation calls were directed at the consorting partner.

However, it is important also to recognize that copulations are more likely to occur outside of the presence of neighboring males in all mating contexts (unguarded, guarded and not contested, and contested) and thus comparisons between female calling frequencies in the presence of neighboring males were based on comparatively low sample sizes. Moving forward it will also be important to assess what drives males and females to avoid copulating while others are nearby and also incorporate more comparisons in female calling behavior based on who is present at the time of copulation

(their relative rank or relative preference versus the current partner). Future playback experiments will be the best way to assess male partner and non-partner “reactions” to female copulation calls. In chacma baboons, Crockford and colleagues (2007) clearly demonstrated that males attend to the calls of females and use those calls to assess the likelihood that females are mating during or outside of consorts. In contrast, Semple

(2001) demonstrated that yellow baboons respond only to calls of consort partners and not non-consort partners (Semple 2001). Although different approaches, these results seem to suggest that males in the two species differ in how much they attend to sounds of copulations in which they are not involved. If so, this would further support differences

101 between the two species in the function of copulation calls. However, replicating similar playback experiments in the two species would better support this hypothesis.

Additionally, no playbacks on olive baboon copulation calls have ever been conducted and their response is predicted to be more similar to yellow baboons.

Female olive baboons call more frequently during consorts than when unguarded.

Similar to the effect of male neighbors on female calling during ejaculatory copulations, this observation may simply reflect the fact that most copulations occurred in the context of consorts, particularly uncontested consorts. Thus, it was not possible to determine if females who call while not guarded are more likely to form a consort with that male later compared to forming consorts with other males who may have been “enticed” into mating as a result of female calling behavior. In future studies, additional observers would be necessary to increase sample sizes of copulations occurring in a variety of contexts, particularly outside of consorts. However, I was able to determine that females who called more frequently had fewer partners and that in dyads where females called more frequently, consorts lasted longer. The correlation between calling frequency and consort length/number of partners suggests calls encourage the consorting partner to continue mate guarding. If copulation calls function to encourage competition, I would have expected females to engage in many short consorts within one cycle, thereby increasing opportunities for sperm competition or for dispersing paternity estimates across multiple male partners (Maestripieri & Roney 2005).

The information that is encoded in female copulation calls in olive baboons also provides some support for the female choice hypothesis. In particular, I demonstrated that

102 females advertise the social status of their current copulatory partners, in terms of both their current rank and stability or trajectory in the dominance hierarchy, by producing calls with more units per second based on higher male social status. If such calls encourage copulating partners to mate guard or discourage opponents to challenge a male of high rank, this may allow females to continue copulating with these “higher-quality” males and bias paternity certainty towards these males. Low-ranking males should avoid disrupting copulations, or competing for access to females, when males are strongly dominant in the hierarchy (Semple et al. 2002; Maestripieri & Roney 2005; Engelhardt et al. 2012). Advertising social status in individual structure also defines copulation call patterns in yellow baboons (Semple et al. 2002) and long-tailed macaques (Engelhardt et al. 2012), although in these species it has been demonstrated that it is strictly rank that influences call structure. I am the first to report the effects of both current rank and rank trajectory on copulation call features.

Support for a new model of copulation calls in primates

Interestingly, results from my study reveal that copulation calls in olive baboons vary in similar ways as long-tailed macaques, one of the only other Old World monkey species in which researchers have as completely studied copulation calls (both frequency and form) in wild groups to assess whether calls vary according to male qualities, female qualities and ovulatory status, and context of the copulation (Engelhardt et al. 2012).

Along with advertising male social qualities, copulation calls in both olive baboons and long-tailed macaques are best predicted by occurrence of male ejaculation and also vary

103 according to the context of the copulation, in particular whether the female is in consort or not (Engelhardt et al. 2012). In long-tailed macaques, though, calls vary acoustically with mate-guarding rather than in their frequency of production. Engelhardt and colleagues (2012) suggest that integrating information on the rank of copulating males, the occurrence of ejaculation, and the context of the copulation (e.g., guarded or not) into copulation calls, both in their production and acoustic features, may support a relatively new model for the evolution of copulation calls developed by Pradhan and colleagues

(2006). This model suggests calls may function to both encourage mate-guarding on the part of high-ranking, or otherwise high-quality males, and additionally attract listener males to mate with females, to reduce the costs of biasing paternity exclusively in one male (Pradhan et al. 2006). In developing this model, Pradhan et al. (2006) have synthesized information from studies that seem to support either the competition/paternity confusion hypothesis or the female choice/mate-guarding.

This model seems particularly appropriate for explaining my results, which generally support the female choice/mate-guarding hypothesis for the evolution of female copulation calls, but also have some features that indicate females call to encourage competition, particularly from higher-quality partners. Evidence for the latter comes from the variation in acoustic features on the basis of both male sociality and female partner preferences. Although producing calls with different acoustic features when mating with males who hold strong positions in the dominance hierarchy should lead weaker males to avoid disrupting copulations, the converse should also hold true: strongly dominant males should attempt to disrupt copulations with weak males. Perhaps with increased sample

104 sizes I will be able to discern whether the calls vary across the ovulatory cycle, with call duration differing according to male quality and time in ovulatory cycle. This could also explain calling and consort patterns in olive baboons. Females seem to call more during uncontested consorts, which could be to encourage high quality partners to extend the consort and further explain the longer consorts among frequent callers. During consorts with males weak in their social status or otherwise eschewed by the female, the calls may encourage better quality males to disrupt the efforts of their current guarding partner.

Thus, as suggested by Pradhan et al. (2006), calls could encourage both guarding, if current partners are high quality, and competition, if current partners are low quality.

If calls function to signal qualities about copulating partners, they meet the criteria associated with “functionally referential” signals (Macedonia & Evans 1993), or calls that “convey specific information about objects or events external to the signaler”

(Engelhardt et al. 2012). I have demonstrated here that olive baboon calls seem to be specific in their production, both in terms of female ovulatory status and male qualities, which is the first criteria of functionally referential signals. The next criteria, that these calls are “interpretable” by receivers of the signals can only be determined through playback experiments.

Playback experimentation has shown that males respond to female copulation calls in ways predicted by hypotheses that suggest these calls reflect a type of post- copulatory sexual selection. Semple and McComb (2000) demonstrated that in Barbary macaques that males respond to calls of females with large sexual swellings compared to females who are in the earlier stages of the estrous cycle. These results indicate that calls

105 are likely salient signals that can be differentiated among male olive baboons as well, but this needs to be directly tested. In particular, it must be demonstrated that males direct attention to calls that are associated with copulations with different types of males, including males who are not as strongly dominant and who may have different levels of association with females.

The patterns that emerged in these calls are relevant particularly as they compare species existing in similar social structures with similar mating patterns. This is important given that Pradhan and colleague’s (2006) model for the dual function of female copulation calls also predicts convergence in calling behavior according to the degree to which males can monopolize females across species. As described previously, calling frequency as well as form is similar across long-tailed macaques, yellow baboons, and olive baboons, especially in terms of their acoustic variation according to male qualities.

These are all species in which males guard females and paternity is biased towards higher ranking males, but there is some relaxation from the priority-of-access model which suggests male rank tightly correlates with male reproductive success (Strum 1982; Smuts

1985; Alberts et al. 2003; Engelhardt et al. 2006). Males in each of these species engage in “following” behavior, where they interact with the consorting pair throughout the consort (Hall & DeVore 1965; Hausfater 1975; Bercovitch 1988; Engelhardt et al. 2006).

Importantly, patterns in yellow and olive baboons seem to be similar to each other and different from those seen in chamcas in ways that are predictable based on subtle but consistent differences in their mating systems (see Introduction). However, acoustic analysis of chacma copulation calls has never been done. Results in all three savanna

106 baboons suggest greater understanding of the complexity of graded signals is possible using a comparative approach.

107

TABLES

Table 4.1: 18 example predictions of the two competing adaptive hypotheses (modified from Nikitopolous et al. 2004 and Maestripieri & Roney 2005).

Female Mate Choice/ Competition/ Encourage Guarding Paternity Confusion Call more frequently near + +/- peak fertility? Call features exaggerated + +/- near peak fertility? Call more frequently with + +/- ejaculation? Call more frequently with - + multiple males? Younger, primiparous, low- +/- + ranking females call more? Decrease time between Dominant/rising rank/ Low-rank/non-preferred, copulations when with? preferred, etc. etc.or regardless of identity Call more frequently when? No audience Audience present

Consort lasts longer? + - Dominant/rising rank/ Low-rank/non-preferred, Call more frequently with? preferred, etc. etc. or regardless of identity Following copulation, - + females solicit other males? Following copulation, - + disassociate with partner? Calls should be more likely Unguarded/Consort Unguarded when? Calling should incite - + disruption of consort Calls should be loud? - +

Who should attend to calls? Partners Non-partners Dominant/rising rank/ Low-rank/non-preferred Dart farther from? preferred, etc. etc. or regardless of identity Low-rank/ non-preferred Calls more exaggerated Dominant/rising rank/ etc. or regardless of (length, rate) with? preferred, etc. identity Calls more exaggerated No audience Audience present (length, rate) when? +/- reflect either neutral or conflicting ideas in the literature

108

Table 4.2: Female identity, group, observation hours and copulation and copulation calling rates for the 19 study females.

Female* Troop Observation Copulation Call frequency time (hrs) rate (cops/hr) (call/ cops) UBG AC 28.68 0.31 0.89 ULY AC 46.17 0.43 0.70 UNK AC 61.38 0.31 0.21 UVA AC 30.93 0.19 0.67 YLT AC 31.55 0.41 0.00 AJA BA 42.15 0.17 0.71 AKA BA 49.68 0.08 0.00 WDF BA 111.28 0.18 0.85 WGR BA 34.97 0.20 0.14 WLD BA 27.65 0.54 0.80 WTW BA 47.67 0.27 0.15 HAU DC 76.92 0.17 0.62 HON DC 54.1 0.54 0.45 HRF DC 52 0.27 0.93 HRS DC 90.29 0.35 0.59 HRT DC 38.98 0.23 0.22 SAS DC 83.62 0.18 0.20 SER DC 54.5 0.18 0.80 SPR DC 72.87 0.36 0.58 *Those with fewer than 10 copulations in which it was clear that they produced a call are italicized.

Table 4.3: Generalized linear mixed models with AICc values that best explain female copulation call production among olive baboon females.

Response Variables β±SE t p AICc Variable included Call frequency Complete -2.563±0.452 -5.666 <0.001 1,144.303 Contested -1.086±0.508 -2.135 0.034

109

Table 4.4: Correlation between copulation call features and female rank and age.

Call feature Rank Age

Total units rs 0.231 0.008 p 0.447 0.979 N 13 13 Call duration rs -0.096 0.196 p 0.754 0.522 N 13 13 Unit rate rs 0.187 -0.196 p 0.541 0.522 N 13 13 Total inhales rs 0.239 0.081 p 0.432 0.791 N 13 13 Total exhales rs 0.318 -0.014 p 0.289 0.964 N 13 13 Exhale rate rs 0.347 -0.430 p 0.246 0.143 N 13 13

110

Table 4.5: Correlation between copulation call features and male characteristics.

Call feature Sociality Experience

Total units rs 0.248 0.238 p 0.498 0.357 N 17 17 Call duration rs 0.040 0.246 p 0.877 0.340 N 17 17 Unit rate rs 0.473 0.287 p 0.055 0.268 N 17 17 Total inhales rs 0.285 0.234 p 0.268 0.365 N 17 17 Total exhales rs 0.161 0.223 p 0.537 0.390 N 17 17 Exhale rate rs 0.349 0.300 p 0.169 0.241 N 17 17

111

Table 4.6: Results from GLMM testing effects of male and female characteristics on a variety of call features across calls from 13 female olive baboons.

Response P Fixed Factor Coefficient SE t value Variable value Partner sociality 0.125 0.161 0.78 0.439 Total units Partner Neutral -0.009 0.275 -0.034 0.973 (inhales a + preferred Eschewed -0.046 0.27 -0.172 0.864 Not exhales) Ovulatory statusb -0.4 0.175 -2.281 0.027 ovulating Partner sociality 0.003 0.043 0.069 0.945 Call Partner Neutral 0.218 0.101 2.164 0.036 a duration preferred Eschewed -0.028 0.097 -0.284 0.777 Not Ovulatory statusb -0.102 0.073 -1.391 0.171 ovulating Partner sociality 0.701 0.412 1.702 0.095 Partner Neutral 0.562 0.953 0.59 0.558 Unit rate a (units/s) preferred Eschewed 1.605 0.927 1.73 0.09 Not Ovulatory statusb -0.315 0.691 -0.455 0.651 ovulating Partner sociality 0.023 0.183 0.126 0.901

Total Partner Neutral 0.168 0.314 0.533 0.597 inhales preferreda Eschewed -0.136 0.323 -0.422 0.675 Not Ovulatory statusb -0.281 0.227 -1.236 0.222 ovulating Partner sociality 0.095 0.112 0.854 0.398 Partner Neutral 0.31 0.273 1.134 0.262 Total a exhales preferred Eschewed 0.324 0.259 1.248 0.218 Not Ovulatory statusb -0.464 0.196 -2.369 0.022 ovulating Partner sociality 0.344 0.227 1.515 0.136 Partner Neutral 0.066 0.525 0.126 0.9 Exhale rate a (exhales/s) preferred Eschewed 0.953 0.503 1.893 0.064 Not Ovulatory statusb -0.19 0.374 -0.511 0.612 ovulating aReference category: preferred bReference category: ovulating

112

Table 4.7: Generalized linear mixed models with AICc values that best explain whether a male contests consorts among olive baboon females.

Response Variable β±SE t value P AICc Variable included value Dart Distance Female Age -0.170±0.039 -4.357 <0.001 406.623 Paritya 0.678±0.239 2.832 0.005 Ovulatingb 0.207±0.083 2.504 0.013 aReference category: Multiparous bReference category: Not ovulating

113

FIGURES Papio anubis

Papio cynocephalus (horizontal arrow indicates 1 second; taken from Semple 2001)

Papio ursinus (Walz and Kitchen, unpublished) data)

Figure 4.1: Comparisons in three copulation calls of three savanna baboon species. Calls from olive baboons (top) last less than 2 s, calls from yellow baboons (middle) average 2.4 seconds and never exceed 6 s (calculated using data from Semple 2001) and calls from chacma baboons typically exceed 6 s (Walz & Kitchen, unpublished data).

114

Figure 4.2: Individual variation in calling during complete copulation in which the male ejaculated or incomplete copulations in which the male did not ejaculate.

Figure 4.3: Individual variation in calling during contested and uncontested consorts. Unguarded not shown.

115

Figure 4.4: Individual variation in calling following complete copulations in the presence or absence of any listener males.

rs = 0.473 p = 0.055

Figure 4.5: Spearman correlation to compare mean call unit rate and male sociality. Points represent the 17 individual males who engaged in copulations with females whose copulation calls were recorded. One male outlier indicated.

116

Figure 4.6: Differences in the number of units between calls expressed by females who are in their ovulatory phase and calls expressed by females outside the ovulatory window (median, first and third quartiles).

Figure 4.7 Differences in the number of exhales between calls expressed by females who are in their ovulatory phase and calls expressed by females outside the ovulatory window (median, first and third quartiles).

117

Figure 4.8: Comparison of the mean call duration (in seconds) between calls expressed while copulating with eschewed partners, neutral partners, and preferred partners. Error bars represent the standard error of the sample mean.

Figure 4.9: Comparison of the mean exhale rate (number per second) of copulation calls expressed with eschewed, neutral, and preferred male partners. Error bars represent the standard error of the sample mean.

118

rs = 0.321 p = 0.040

Figure 4.10: Spearman correlation to compare calling frequency and number of days spent in consort. Dots represent the 41 male-female dyads who engaged in copulations during consorts in this data set.

1.2

0.8

0.6 y = -0.29x + 1.24

0.4

0.2

Copulation Calling Frequency CallingCopulation 0 0 1 2 3 4 5 Average Number of Male Partners

Figure 4.11: Copulation calling frequency versus average number of male partners per female.

119

p < 0.05

Figure 4.12: Comparison of the mean + SE dart distance (in meters) of females following complete copulations (in which a male ejaculated) and incomplete copulations. Means were averaged across individual female means.

Figure 4.13: Mean dart distances (in meters) following completed copulations across females of different ages.

120

p < 0.05

Figure 4.14: Comparison of the mean + SE dart distance (in meters) of females during copulations that took place during different phases of the cycle. Means were averaged across individual female means.

121

CHAPTER 5: DO FEMALE OLIVE BABOONS DIRECTLY COMPETE FOR ACCESS TO MATES?

Introduction

Throughout this dissertation, I have primarily considered the intersections and effects of male-male competition, male coercion, and female mate preferences on male- female interactions and copulatory success. This is reflective of the broader primatological literature on mechanisms of sexual selection, which has increasingly focused on female reproductive strategies as counterstrategies to male coercion or competition (Setchell & Kappeler 2003) and, more generally, on sexual conflict

(Palombit 2014). Data presented in Chapters 3 and 4 largely conform to this perspective, suggesting that the behavioral mechanisms indicative of mate “choice” and male competition interact in complex ways that can impact the likelihood that certain male- female dyads will interact and potentially reproduce. However, female-female competition is an additional, though less widely considered, mechanism of selection that may affect the likelihood that certain male-female dyads interact. Thus, female-female competition has the potential to prevent females from expressing preferences for specific male partners (Zumpe & Michael 1985; Setchell & Kappeler 2003; Clutton-Brock 2009;

Huchard & Cowlishaw 2011). In this chapter, I consider whether female olive baboons directly compete for access to male partners and whether this is influenced by female partner preferences in particular.

122

Factors influencing female survivability and ability to reproduce successfully, including accessing and defending high-quality food sources, have been consistently demonstrated to influence female reproductive success, contributing to the social rank typical of multifemale primate societies (Fedigan 1983; van Noordwijk & van Schaik

1987; Barton & Whiten 1993; Altmann & Alberts 2003; Pusey 2012). However, inability to mate (or to mate with preferred partners) has obvious consequences on female fitness as well and the influence female competitors have on likelihood of successfully mating should not be overlooked (Zumpe & Michael 1985). This is true even in species that exemplify “conventional” sex roles with male-biased operational sex ratios (OSRs)

(Huchard & Cowlishaw 2011) and behavioral and morphological indicators of intense male-male competition (i.e., direct vocal and behavioral competition and pronounced sexual dimorphism). Researchers have recently theorized that “sex roles” in these species may be more flexible than has been traditionally realized, and males and females may both be selective in their mate choices and competitive in vying for access to mates

(Clutton-Brock 2009; Gowaty & Hubbell 2009). As Huchard and Cowlishaw (2011) suggest, it is important to broaden our “traditional” perspectives on intrasexual and intersexual mechanisms of sexual selection.

Of those primatological studies focusing on female intrasexual competition in species with male-biased OSRs, most have focused on species that are less intensely male-biased, where competition among females for mates is expected to be more common. For example, in uni-male, “polygynous” systems, access to male partners may be more constrained than in multi-male societies. Studies indicate females, particularly

123 dominant individuals, limit mating opportunities for subordinate females through directed aggression and harassment in these species (Erythrocebus patas: Loy & Loy 1977;

Theropithecus gelada: Dunbar 1984; Presbytis entellus: Sommer 1989; Gorilla gorilla:

Watts 1990).

Data on female-female harassment as it occurs in the context of mating in multimale, multi-female primate societies in which females mate polygynandrously are even more limited. Most of those few studies conducted tend to focus on seasonal breeders, which are similar to species characterized by uni-male systems because the OSR bias towards males is relaxed. Nevertheless, these studies are informative and demonstrate effects of breeding competition on female behavior and reproductive physiology. For example, data on a free-ranging population of rhesus macaques (Macaca mulatta), a seasonal breeder, indicate that high-ranking females disrupt copulations of low-ranking females (Loy 1971). In species like Hanuman langurs (Presbytis entellus), which can both exist in one-male social groups and breed seasonally, females express “estrus” following conception, which is consequently non-conceptive but can still result in copulations with males, limiting the sperm available for conception with other females

(Small 1988). In ring-tailed lemurs (Lemur catta), female-female conflict rates are higher during mating seasons than outside of the mating period, similarly indicating competition for sperm characterizes this seasonally breeding species (von Engelhardt et al. 2000).

The notion that males, and consequently sperm, are less of a limited resource in non-seasonal, multi-male species with OSRs that are clearly male-biased has likely limited researchers from testing the effects of female competition for mates in these

124 species. However, Huchard and Cowlishaw (2011) argue savanna baboons are a quintessential mammalian model in which to test direct competition as it occurs among females in species with “typical” sex roles: females express prolonged periods of gestation and lactation and males compete intensely for reproductive access to females.

Nevertheless, females may also compete heavily in a variety of contexts that affect reproductive success. In particular, there is evidence for competition over potential infant caregivers (Palombit 2001). “Friendships” (defined by Smuts 1985) between male and female baboons across all baboon species can increase infant survivorship through protection from lethal or non-lethal harassment primarily from male and female conspecifics (Busse & Hamilton 1981; Smuts 1985; Strum 1987; Buchan et al. 2003).

Palombit and colleagues (2001) have demonstrated that high-ranking female chacma baboons (Papio ursinus) can “displace” low-ranking females from their friendships with shared male friends. Low-ranking females must seek out alternative strategies to derive the benefits that having male friends affords, particularly when infants are young and vulnerable.

Researchers have considered the impact of indirect reproductive competition in baboons over caregivers (and food: see Bercovitch and Strum 1993; Packer et al. 1995); however, the question still remains: do female baboons directly compete for mating opportunities? Although breeding is nonseasonal and females are not synchronously receptive in baboons, males consort with receptive females (with receptivity indicated by their large sexual swellings: Higham et al. 2008; Huchard et al. 2009) during prolonged periods making individual males (including potentially preferred males) unavailable as a

125 potential mating partner for other females who may be sexually receptive at an overlapping time (Huchard et al. 2009). These restrictions on availability of male partners are similar to contexts that restrict male partners and sperm in more seasonally breeding species and those who exist in uni-male systems. Thus, direct competition should characterize baboon societies (Huchard & Cowlishaw 2011). Some limited testing of this hypothesis has been done. Huchard and Cowlishaw (2011) demonstrated that in chacma baboons, pregnant females attempt to limit conception among cycling females, directing high rates of aggression towards these females. Similarly, in yellow baboons (Papio cynocephalus), females, particularly pregnant females, direct high rates of aggression towards cycling females, females in their final phase of pregnancy, and those who have most recently given birth (Wasser & Starling 1988). These patterns may reflect attempts to suppress reproduction in female competitors and thereby reduce the number of future feeding competitors (Huchard & Cowlishaw 2011). Here I expand this work to include the third savanna baboon, olive baboons (Papio anubis), never previously tested for female-female competition surrounding mating.

Although similar to their close phylogenetic relatives the chacma baboons in that most conceptive copulations occur in the context of consorts, olive and yellow baboons differ in that non-consorting males frequently follow consorting pairs for prolonged periods of times, even if there are other receptive females in the group (Hall & DeVore

1965; Bercovitch 1988; Danish & Palombit 2014). This “satellite” strategy (Gross 1996) by challenger males in yellow and olive baboons creates a context in which the number

126 of viable male partners is further limited for non-consorting females, and should increase competition around mating, particularly among cycling females.

Here, I explore competition among cycling olive baboon females based on social and reproductive contexts. I hypothesized rates of aggression exchanged will be most intense when females are: (1) ovulating, (2) consorting, (3) guarded by a preferred partner, and (4) followed by multiple males. I also make several hypotheses about the reproductive state of female competitors. If female competition occurs directly for access to mates, I hypothesize cycling females will exchange (direct and receive) the highest rates of aggression with other cycling females and that aggression overall will be higher when there are higher proportions of cycling females in the group. Alternatively, if cycling females are competing to increase their likelihood of establishing friendships by biasing paternity towards specific males, than cycling females should direct the most aggression towards lactating and/or pregnant females, who are the most likely to interfere with attempts at early bond formation. If competition over mates occurs through reproductive suppression, which seems to be the common pattern among chacma

(Huchard & Cowlishaw 2011) and yellow baboons (Wasser & Starling 1988), I hypothesize cycling olive females will receive the most aggression from pregnant females. Finally, I hypothesize that high ranking females will give the most and receive the least aggression and that rates of aggression exchanged by females will be most intense between nonkin and those who are more socially distant (i.e., with greater disparities in social rank, indicative of weak social bonds in baboons: Kapsalis 2003; Silk et al. 2006; Silk et al. 2010).

127

Results

General Results

Across all focal observations (N=666, by 19 females over 1035 hours), females exchanged 676 agonistic interactions with other females in their social groups. Focal cycling females received more than double the amount of aggression (mean: 0.50 interactions/hour, SD = 0.36) than they directed (mean: 0.24; SD = 0.17). There was considerable variation across individuals in both rates of aggression received (mean +/-

SD: from 0.05 +/- 0.18 to 0.78 +/- 0.48/hr) and directed (mean +/- SD: 0.25 +/- 0.11 to

0.72 +/- 0.38/hr). Below I examine what specific factors might explain this variation.

Does the proportion of cycling females in the group influence overall rates of aggression?

I ran a generalized linear mixed model (GLMM) (which allows me to control for individual contribution of multiple points to the data set by including female identity nested in group identity as a random effect; see Chapter 2) to investigate whether the proportion of cycling females in groups influenced aggression rates. I found rates of female aggression exchanged were not influenced by the proportion of cycling females

(GLMM: β±SE = 0.853 + 2.129, t = 0.401, p = 0.689).

Do cycling females receive or direct more aggression during different social and reproductive contexts?

As expected, dominance rank of the focal female negatively correlated with rates of aggression received (Spearman correlation: rs=-0.478, N=19, p=0.038) and positively correlated with rates of aggression given or “directed” (rs=0.586, N=19, p=0.008).

128

Demonstrating these clear relationships with rank, similar to results in other studies (e.g.,

Smuts 1985), provides an internal validation of my research. However, there was no correlation between female experience (a composite index where higher values represents older females with more offspring birthed; see Chapter 2) and either rates of aggression received (rs=0.006, N=19, p=0.978) or rates of aggression directed (rs=0.074, N=19, p=0.750).

Across all focal follows (N=666), the presence of a consorting male also contributed to rates of aggression females received and directed, but in the opposite direction as I hypothesized. Females exchanged higher rates of aggression outside of consorts than during consorts (Mann-Whitney U test, directed U= -2.615, p=0.009; received: U= -7.438, p<0.001). Perhaps this is because the presence of a consorting male partner limits the ability of other females to get close and either harass or be harassed by his mating partner. Low-ranking females seem to benefit from this relationship most

(Figure 5.1). The rates of aggression these females received during consorts were similar to rates among higher-ranking females and were much lower than rates they experienced outside of consorts. The only exceptions were “alpha” females who actually faced slightly more aggression in consorts than outside consorts.

Although most female-female aggression occurred outside of consorts, it is important to note that not all ovulating females were involved in a consort (and not all consorts were with ovulating females). When including these non-consorting females, female ovulatory status predicted rates of aggressive interactions exchanged. Cycling focal females received higher rates of aggression during ovulatory days (Mann-Whitney

129

U=-3.261, p=0.001). Conversely, the focal female’s ovulatory status did not impact rates of aggression she directed (U=1.057, p=0.290). This suggests that these females with highly visible, large sexual swellings were targeted when they were unguarded.

Furthermore, these results suggest it is not simply because there were certain days when all females were highly aggressive as the targeted female did not also direct more aggression on those ovulatory days.

Finally, the reproductive status (i.e., cycling, pregnant, or lactating) of the competitor female did not predict the rates of aggression focal cycling females received

(Mann-Whitney U test, cycling vs. lactating: U=-1.314, p=0.189; cycling vs. pregnant:

U=-0.953, p=0.340; pregnant vs. lactating: U=0.445, p=0.656), but did predict the rates of aggression focal females directed. Focal cycling females directed higher rates of aggression towards lactating than towards other cycling females (U= 2.049, p=0.040) or towards pregnant females, although this latter result only approached statistical significance (U= 1.775, p=0.076; Figure 5.2). There was no difference between aggression directed at cycling vs. pregnant (U=0.076, p=0.939)

Surprisingly, and contrary to my predictions, whether the competitor female was a close kin of the focal female did not predict rates of aggression received (Mann-Whitney

U= 0.791, p=0.429) or directed (U= -1.196, p=0.232). Likewise, social distance between competitor and focal female did not correlate with rates of aggression received

(Spearman correlation, rs=0.026, p=0.534) or directed (rs=0.010, p=0.559).

Using a GLMM allowed me to compare relative effects of all above variables

(with female identity nested within group identity as a random factor). Using the

130 predictors described above (i.e., social rank, experience, kinship, social distance, reproductive status of the competitor, ovulatory and consort status of the focal female), this GLMM confirmed some of the above results: in the best reduced model (based on lowest AICc; see Chapter 2) social rank of the focal female and consort status were retained as primary predictors of rates of aggression received and directed between females across the study (Table 5.1; Figure 5.1).

Do rates of aggression change according to the context of the consort?

It was clear that most aggression between females occurred outside of consorts.

However, if females directly compete for access to mates, they likely direct aggression differently while in consorts and thus the consort context should be associated with different predictions than across the entire ovulatory cycle. Therefore, I restricted the analysis to the subset of focal samples occurring during consorts (N=162 follows on 19 females) and tested whether rates of female-female aggressive interactions were related to preferences (i.e., preferred, neutral, eschewed) of the consorting male, whether the consort was contested, whether the female was ovulating, and the reproductive status of the competitor female.

During consorts, females directed higher rates of aggression towards competitor females when in consorts with preferred partners than with eschewed partners (Mann-

Whitney U= 2.301, p=0.021). Rates of aggression received from competitor females were lower while in consort with preferred partners compared to eschewed partners

(Figure 5.3), but this result only approached statistical significance (U= -1.799, p=0.072).

131

Partner preference had no other effect on rates of aggression received (eschewed/neutral:

U=-1.212, p=0.225 neutral vs. preferred: U=-0.088, p=0.930) or directed (eschewed vs. neutral: U=1.198, p=0.231; neutral vs. preferred: U=0.310, p=0.757).

As with data from across the entire cycle, analysis of the consort subset revealed focal cycling females directed higher rates of aggression towards lactating than other cycling females (Mann-Whitney U=2.301, p=0.021) or than pregnant females (U= -

3.136, p = 0.002). Rates of aggression directed did not differ between cycling and pregnant females (U=-1.077, p=0.281) and rates of aggression received did not vary with the reproductive status of the competitor female (lactating vs. cycling: U=-1.566, p=0.117; lactating vs. pregnant: U=1.597, p=0.110; cycling vs. pregnant: U=-0.099, p=0.921).

Females directed higher rates of aggression during consorts while ovulating

(Mann-Whitney U= 2.758, p=0.006), but did not receive higher rates of aggression during this time (U= -1.685, p=0.092). This is a different direction for the results than when unguarded females are included (above section). This suggests that only when ovulating females have a partner (i.e., in consort) are they aggressive to other females and that the presence of the male in this guarded situation may protect the female from receiving more aggression.

Finally, contrary to hypotheses, the presence of a contesting male did not predict rates of aggression directed (Mann-Whitney U = 1.595, p=0.111) or received (U = -

1.566, p=0.117).

132

Of the predictors I considered here, there seemed to primarily be effects on aggression directed rather than received. Thus, I tested the relative effects of all of the above predictors on rates of aggression directed using a GLMM (with female identity nested in group identity as a random effect). In this analysis, the best model retained the reproductive state of the competitor female and the focal females’ ovulatory status as primary predictors of rates of aggression females directed during consorts (Table 5.2;

Figure 5.4).

Do rates of female-female affiliations differ generally across different reproductive contexts?

Contrary to my hypotheses, neither relatedness or social distance predicted rates of aggression that females directed or received during or outside of consorts (see above), suggesting these variables do not influence agonistic interactions in olive baboons. It may be that kin and nonkin are less likely to come into contact and thus less likely to direct aggression towards one another during different mating contexts. Therefore, I also tested whether Affiliation Indices (see Chapter 2) varied among kin and nonkin depending on whether or not focal females were engaged in consorts. An Affiliation Index close to 0 indicates females in a dyad spend little time grooming and maintaining close proximity.

Among nonkin dyads (N=246), indices were significantly lower while females were in consorts than when not in consorts (Wilcoxon signed-rank test: T = 4.261, p<0.001), but indices were not significantly different during consorts vs. outside of consorts among kin dyads (N=35; T = -0.413, p=0.679; Figure 5.5). Thus, immediate kin interact at similar

133 rates regardless of their mate-guarding status, while nonkin interact less often when females are being mate-guarded.

Discussion

Overall, more aggression was exchanged with a cycling female when she was unguarded and, during this time, cycling female olive baboons received higher rates of aggression than they directed, particularly during ovulatory days. These results are somewhat consistent with studies investigating aggression exchanged between females around reproduction in both chacma baboons (Huchard & Cowlishaw 2011) and yellow baboons (Wasser & Starling 1988), which indicate that cycling females, especially females expressing pronounced sexual swellings, are targeted by females.

However, a primary difference between my study and results from Huchard and

Cowlishaw’s (2011) study on female competition in chacma baboons is that they found that swollen females experience higher rates of aggression when in consorts compared to when unguarded. In chacma baboons, female competitors follow the dyad and “harass” the consorting female (Huchard & Cowlishaw 2011). Cheney and Seyfarth (2007) reported two cases in which females were able to completely disrupt consortships, ending them sooner than expected based on female swelling size, including one case where the disrupting females was lower-ranking than the consorting female. This following and harassing pattern is not common in olive baboons (pers. obs.). Furthermore, in studies on yellow and chacma baboons, it is the pregnant females who direct these high rates of aggression towards cycling females; thus, researchers suggest this is an attempt on the

134 part of pregnant females to suppress reproduction among cycling females to avoid competition over resources in the future (Wasser & Starling 1988; Huchard & Cowlishaw

2011). However, in my study, cycling females did not receive significantly different rates of aggression from different females (cycling, pregnant, or lactating) whether we looked at only during consorts or across the entire study. This targeting of swollen females in consort by pregnant females suggest that perhaps the ability of female competition to suppress reproduction is stronger in chacmas and yellow than in olives. It is unclear why yellow baboons should be more like chacma baboons in this aspect. More data on birth patterns would be needed to test whether reproductive suppression strategies succeed.

Perhaps there is another explanation. Among chacma baboons, Huchard and

Cowlishaw (2011) suggest that, rather than patterns of agonism indicating this is a result of direct competition over mates, it instead reflects patterns of incidental aggression that females receive because they are shifting their normal patterns of movement and associating more closely with females they might not otherwise interact with, especially high-ranking females. As in chacma baboons, olive baboons may shift patterns of movement during consorts. However, unlike in chacma baboons, who appear to associate more closely with females they might not otherwise and thus are susceptible to increased harassment from other females (Huchard and Cowlishaw 2011), consorting female olive baboons may avoid the most aggressive females. Some of my data support this in that olive baboon females affiliate less with nonkin, who are more likely to harass females generally (Boese 1975; Seyfarth 1977), but maintain similar association patterns with

135 immediate kin whether in or out of consort. In the future, I will analyze other social data I collected to evaluate this.

Additionally, lower aggression exchanged during consorts in olive baboons could simply be a result of having a consorting male present who “thwarts” agonism from particularly aggressive females and perhaps this is not the case in chacmas. In yellow baboons, Hausfater (1975) demonstrated that fully swollen females received less aggression from other females during consorts, in part, because consort partners used herding behavior to guide them away from possible aggressive interactions. Male olive baboons also use herding to guide females during consorts (Bercovitch 1995) and, thus, a similar pattern may be expressed in olive baboons. Though less tolerant of agonistic interactions, Smuts (1985) reports that consorting males are relatively tolerant of affiliative female-female interactions during consorts. Similarly, I found that low-ranking females, particularly those in the bottom 50% of the dominance hierarchies, seem to benefit from this association most. The rates of aggression they received during consorts were similar to rates among higher-ranking females and were much lower than rates they experienced outside of consorts. Moving forward, it will be important to assess whether females gain either social or reproductive benefits in associating with close kin while those females are being followed by a male, and also whether reduced harassment from females received during consorts benefit females proximally (e.g.,as reduced stress, measured through glucocorticoid hormones: Sapolsky 1983, Beehner & Whiten 2004).

It may also be that (along with disassociating from aggressive females and being protected to some degree by consort partners), males and females who are engaged in

136 consorts spend more time away from the group more generally in olives and not chacmas.

Although I did not directly test whether females’ spatial movements changed during and outside of consorts, anecdotally, there were several cases where females and consorting males would be absent from the group for much of the day, or would be located ranging significantly farther away from the main group. Future research could compare the ranging patterns behind when females and males are in consorts versus not in consorts using GPS data collection or radio tracking devices combined with point pattern analysis

(Wentz et al. 2003).

As described above, cycling olive females were not targeted by pregnant females and thus the function of female-female aggression in a mating context does not seem to be reproductive suppression. However, results also indicate that direct mating competition is not the primary driver of female aggression in this species. Although unguarded cycling females were targeted, they were not targeted disproportionately by other cycling females, whether I looked across the whole study or just during consorts.

Also, rates of aggression were not highest during periods when a higher proportion of females were cycling in the group, which is opposite the pattern seen in yellow (Wasser

& Starling 1988) and chacma baboons (Huchard & Cowlishaw 2011) as well as other seasonal breeding species living in multi-male, multi-female societies (e.g., ring-tailed lemurs: von Engelhardt et al. 2000). Also, if direct competition for mates is the typical pattern in this species, we would expect females to receive higher rates of aggression during consorts, particularly from other cycling females. I revealed the opposite pattern, where cycling females both received and directed aggression more often when

137 unguarded. Although rates of aggression are generally lower during consorts, this is the most important time for females to conceive so I examined this smaller data subset separately. I found that when guarded, cycling females directed more aggression than they received. The lack of received aggression might be due to the presence of the male, as I explored above. However, other patterns revealed during consorts suggested the function of aggression might be different than outside of consorts. First, cycling females in a consort are more aggressive during days surrounding ovulation, suggesting that the presence of a partner makes them more aggressive near conception. During this time, it would seem that females attempt to limit competitors from disrupting their mating attempts.

Although this pattern could be explained as direct competition for a mate if females targeted other cycling females, I found they instead targeted lactating females.

Evidence presented in Chapter 4 suggests that females may attempt to maintain consortships with certain males to bias paternity towards those males. I found that these include males who are ascending the dominance hierarchy and those who tend to remain in the group for extended times, which are qualities that may be indicative of good future

“friends” for females and their infants (Smuts 1985). Relationships between males and their female friends’ infants often last until infants are three or four years old (Johnson

1984; Smuts 1985). Because there is some evidence that these friendships begin when males and females copulate frequently during conceptive cycles (Bercovitch 1995), lactating females who already have dependent offspring may be perceived as the greatest

138 threat to disrupting the forging of these bonds among potentially conceptive, consorting females.

This is similar to patterns expressed among female yellow baboons, who direct high rates of aggression towards mothers shortly after they have given birth (Wasser &

Starling 1988), but quite opposite the pattern reflected in chacma baboons, who have been show to direct low rates of aggression towards lactating females compared to females in other reproductive states (Huchard & Cowlishaw 2011). These patterns may reflect differences in the function of friendships in olive and yellow baboons versus chacma baboons. In all three species, there is evidence that friendships begin during conceptive cycles and thus the male friend is at least more confident in their likelihood of having sired the infant (Bercovitch 1995; Palombit et al. 1997; Moscovice et al. 2009).

For chacma baboons, these friendships seem to serve as primarily a counterstrategy to male infanticide (Busse & Hamilton 1981) and dissolve with the death of the infant or following their most vulnerable developmental periods (Palombit et al. 1997; Moscovice et al. 2009). Among olive baboons and yellow baboons, in which infanticide seems to be a less common male reproductive strategy than among chacmas (Palombit 2003), males remain friends with females for several years beyond the weaning period, and bond with and protect juveniles from predators (see Chapter 3) and other juveniles and protect females from nonlethal harassment from conspecifics, including harassment from other females (Ransom & Ransom 1971; Lemasson et al. 2008), a behavior that is never observed in chacma baboons (Kitchen, pers. comm.). Thus, it makes sense that in olive baboons and yellow baboons, females may attempt to disrupt these friendships in order to

139 increase the likelihood that their current mating partner will be able to protect them and their offspring in the future. Because I did not follow all females in all reproductive states, I could not directly test whether those lactating females who cycling females were more aggressive towards were the current friends of consort partners. This is an important question to address in future studies both at Gombe and in other olive and yellow baboon populations (see Chapter 6).

Female olive baboons also directed higher rates of aggression during consorts with preferred compared to eschewed partners. This further lends support to the idea that directing aggression around ovulation, especially towards lactating females, when in consorts is part of a strategy that enables females to avoid disruptions of bonding with potential sires of offspring. It is unclear why partner preference was not a stronger predictor in models testing relative effects of directed aggression, and thus it will be important to continue testing the relationship between female partner preference and female-female aggression to determine if females preferentially direct agonism towards other female to increase the likelihood of successfully copulating with males whom they more frequently solicit.

Based on this and other studies, direct competition for mates does not seem to seem to be the primary function of female-female agonism in South and East African savanna baboon species. In both chacma and yellow baboons there seems to be some support for a role of reproductive suppression, whereas in both yellow and olive baboons there seems to be evidence of indirect competition for friends. However, females may benefit from indirect competition with conspecific females if it increases their ability to

140 form long-term bonds with males who increase survival of offspring. My data suggest that this strategy seems to be at work in olive baboons as females were most aggressive during a consort when they were closest to ovulation, when the other female was lactating (lactating females also form “friendships” with males), and when in consorts with preferred partners. To better assess whether indirect competition increases female reproductive success, it is important to accumulate long-term data on male-female friendships among the Gombe baboons and assess male paternity of offspring at this site.

I address this as an important direction of future research in Chapter 6.

141

TABLES

Table 5.1: Generalized linear mixed models with AICc that best explain rates of aggression received and directed by cycling olive baboon females.

Response Variables β±SE t p AICc Variable included Aggression Social rank -0.765±0.215 -3.563 <0.001 1,437.608 received Not 0.259±0.104 2.496 0.013 guarded

Aggression Social rank 1.429±0.374 3.825 <0.001 2,156.453 directed Not 0.456±0.168 2.708 0.007 guarded

Table 5.2: Generalized linear mixed models with AICc values that best explain rates of aggression directed by cycling olive baboon females in consorts.

Variables β±SE t p AICc included Ovulatory Not ovulating -0.643±0.296 -2.176 0.031 662.812 statusa Cycling 0.587±0.397 1.476 0.142 Reproductive Lactating 1.014±0.416 2.435 0.016 statusb aReference category: ovulating bReference category: pregnant

142

FIGURES

(a)

(b)

Figure 5.1: Comparison of the mean rate (number per hour) of aggression (a) received and (b) directed among cycling females based on rank (with higher ranks being higher numbers; see Chapter 2). Data points represent individual females in consort (blue) or outside of consorts (green).

143

(a)

NS NS NS

(b)

p= 0.040* NS p = 0.076

Figure 5.2: Comparison of the mean rate (number per hour) of aggression that focal cycling females (a) received and (b) directed towards other females of varying reproductive states. Error bars represent the standard error of the sample mean. “NS” identifies nonsignificant relationships at p>0.100. “*” identifies significant relationships at the α=0.05 level.

144

(a)

p=0.072 NS NS

(b)

p = 0.021* NS NS

Figure 5.3: Comparison of the mean rates (number per hour) of aggression focal cycling females either (a) directed or (b) received while in consorts with partners differing according to female preferences. Error bars represent the standard error of the sample mean. “NS” identifies nonsignificant relationships of p > 0.100. “*” identifies significant relationships at the α=0.05 level.

145

146

(a)

(b)

Figure 5.5: Comparisons of mean Affiliation Indices while females are both guarded and not guarded among (a) nonkin and (b) kin. Error bars represent the standard error of the sample mean. “NS” identifies nonsignificant relationships and “*” identifies significant relationships at the α=0.05 level.

147

CHAPTER 6: CONCLUSIONS

Summary and Contributions

My study adds to a small but growing body of research demonstrating possible interacting effects of female preference, female-female competition, and male aggression in primates (reviewed by Kappeler 2012). Specifically, my study revitalizes the body of literature on the Gombe olive baboon (Papio anubis) population (Packer 1979, 1980;

Ransom 1981; Packer et al. 1995) using new methods in the study of female reproductive physiology, social and sexual behavior, and sexual conflict theory. Importantly, mine is the first study to comprehensively consider possible predictors of copulation calling behavior in a wild group of olive baboons. Because copulation calls are vocalizations that likely evolved in catarrhine primates as a result of interacting mechanisms of sexual selection, incorporating tests of the possible function of this sexual signal into studies addressing the co-evolution of female and male strategies will be important to advances in sexual conflict theory moving forward (Palombit 2014).

My study revealed that the social and sexual lives of female olive baboons differ as they approach periods in their ovulatory cycles, when they are more likely to conceive.

This is not a completely new finding as other studies have also demonstrated this both in captive and wild populations (Manson 1992; Soltis 1999b; Soltis et al. 2001;

Nikitopoulos et al. 2005; Stumpf & Boesch 2005, 2006), although few have looked as

148 comprehensively at the problem as I have here by considering differences in pre- and post-copulatory behaviors throughout the ovulatory cycle as well as the role female competition plays in these behavioral strategies in a wild group of nonhuman primates.

Around ovulation, females produce copulation calls with a faster unit rate, have shorter dart distances following copulation, are more aggressive to other females, copulate more with preferred males and with males ascending the hierarchy, and have no propensity to copulate based on direct male aggression. My study supports hypotheses suggesting female strategies, like darting, copulation calls, and proceptive behaviors toward males with certain qualities appear to serve as a graded signal in the way that Nunn (1999) suggests sexual swellings function (van Schaik et al. 2004; Clarke et al. 2009). Because females produce some of these calls and behaviors outside of ovulation but the nature of them change closer to ovulation, they can be combined with sexual swellings to both confuse and confirm paternity (Nunn 1999). Thus, my study suggests further complexity in the graded nature of sexual signals and in what constitutes a sexual signal.

By acting differently when outside versus inside the ovulatory window, females can enlist both paternity confusion/male competition and female choice/mate-guarding strategies. Although most of my results on copulation calls suggest a female choice/mate guarding role (calling more when no male listeners besides the partner were nearby, with calling frequency negatively correlated with number of partners and positively correlated with duration of the consort) there were some features that might be used to elicit competition such as the rate that units with the call were produced and perhaps aspects of darting behavior. Outside the ovulatory window females also behaved differently – they

149 were more likely to copulate with males based on male rank and male aggression toward them. My results are comparable to a recent study investigating dual function in both the expression and form of copulation calls in long-tailed macaques (Engelhardt et al. 2012).

Thus, copulation calls in olive baboons, and other species in which male competition for mates is intense but females still have some ability to manipulate who mates, calls may both function to incite competition/confuse paternity and encourage mate-guarding from different types of males.

Evidence presented here, particularly from darting behavior, also indicates that post-copulatory strategies may differ across females of different ages and reproductive histories. For younger, nulliparous females, who likely have less experience with male coercion and developing bonds with males (Smuts 1985), it may be most advantageous to recruit as many males to compete for copulations and potentially serve as friends in the future. This is similar to results from a recent study on copulation calls produced in chimpanzees (Fallon et al. 2016), which indicates young, nulliparous females produce copulation calls less discriminately than do older, parous females. Thus, post-copulatory sexual signals may differ within females according to their fertility status as well as between females according to their reproductive histories.

In addition to males ascending the hierarchy and males with longer tenures having some increased success with females near the ovulation window, males who were classified as “preferred” also had these advantages. However, it remains unclear what females are using to determine whether a male is preferred or not. One characteristic potentially driving female mate preferences are males who will act as a friend and

150 protector of offspring. Although infanticide is rare in Gombe baboons (Palombit et al.

2000), females and their infants can be brutally attacked by males (Smuts 1985).

Additionally, chimpanzees can be predators of baboon infants at Gombe (Wrangham &

Riss 1990) and their presence may be an additional selective pressure among females for males who can protect vulnerable offspring. There is some evidence that friendship is a motivator for females in that the target of cycling female aggression near ovulation seems to be lactating females. These are the females who would be vying for attention from males as friends and thus would be the biggest competitors for cycling females in terms of time taken from current mating opportunities. Furthermore, my study reveals that, while perhaps a potential limiter in their ability to enact choice, consorts may also provide females protection from female competitors, especially among young, low- ranking females, who are frequent targets of aggression by conspecifics (Smuts 1985).

One thing that does not seem to be a predictor of copulation near ovulation is direct coercion. Bercovitch (1995) first predicted this from his research on olive baboons at Gilgil, Kenya and my data corroborate his hypothesis. Unlike among chimpanzees (one of the few other species to be consistently studied to try to disentangle male coercion from female choice in primates) (Stumpf & Boesch 2005, 2006; Muller et al. 2011), I found no evidence that preferred males were more aggressive near ovulation or when looking across the females entire cycle. Additionally, while extreme sexual dimorphism characterizes this species and males can be physically aggressive towards females throughout the ovulatory cycle in ways that can cause substantial injuries (reviewed in

Kitchen et al. 2009a) and may influence their preferences; direct coercion does not appear

151 to be an important reproductive strategy for males in olive baboons. One important alternative strategy for males seems to be contesting a consort and female choice maybe the primary driver of these strategies given that males who are ultimately successful in taking over a consort are those who are “preferred” by the female.

I have also made strides in demonstrating important similarities and differences within the Papio genus, especially between East African yellow and olive baboons and their South African relatives, the chacma baboons. This is exciting as recent molecular evidence suggests that the southern clade of chacma baboons is definitely distinct and the oldest group (Zinner et al. 2013). Overall, especially in terms of patterns of copulatory signals and female competition, female behavior to bias paternity and avoid coercion and aggression from conspecifics seems to be a driver of the long, protracted friendships that both males and females can benefit from in olive and yellow baboons (Johnson 1984;

Smuts 1985; Lemasson et al. 2008), whereas female behavior to confuse paternity drives the behavior of chacma baboons (and thus limit infanticide which is the leading cause of infant death only in chacma baboons: Palombit et al. 2000), who engage in short-term friendships that primarily function to protect against coercion in the form of infanticide

(Busse & Hamilton 1981; Palombit et al. 1997; Moscovice et al. 2009). Results in all three savanna baboons suggest greater understanding of the complexity of graded behavioral, physiological, and vocal signaling is possible using a comparative approach.

152

Limitations

While I was able to make significant contributions to the study of female choice and intersecting mechanisms of competition and coercion on female preferences in olive baboons, there were some limitations to my study that prevented me from completely and fully answering all the questions I set out to explore. In particular, I was unable to assess whether females maintain relationships with preferred partners as friendships during vulnerable periods around infant development primarily because, of the four females who conceived during my first study period, two infants died within the first three weeks of life. Following the second study period, my field assistant was unable to continue collecting behavioral and genetic data during the necessary period after infants were born. In the future, I hope to continue work begun here, and assess through longitudinal studies whether preferred partners or those males who females copulated with and signaled receptivity towards are more likely to be friends (see Future Directions below).

Furthermore, olive baboon groups at Gombe are relatively small, and thus, to have a consistent sample of cycling females from which to obtain behavioral, vocal, and hormonal data, it was necessary for me to focal females from three different groups. This required that I invest extra time moving between groups and also impacted my ability to collect fecal samples on a daily basis from all focal subjects. Thus, my estimation of the conceptive period is broader than is ideal when using fecal hormone samples to assess female reproductive cycling. However, my method was still preferable to relying exclusively on sexual swelling size and date of detumescence because results from other olive baboon populations indicate ovulation can happen outside the full swelling period

153 in this species (Higham et al. 2008). A possible solution to this problem used in other primate studies (Deschner et al. 2004; Higham et al. 2008; Engelhardt et al. 2012) has been to assay fecal samples for both fecal estrogen and progestogen (PdG) concentrations, which together can provide a more narrow ovulatory window in olive baboons (Higham et al. 2008). However, PdG assays were not possible in my study because this hormone requires that C12 cartridges be washed and fixed with sodium azide (Beehner, pers. comm.), which I could not access in the field. Furthermore, Higham and colleagues (2007) indicate that consumption of the African black plum (Vitex doniana) and related Vitex species can influence the expression of PdG in olive baboon populations that consume this fruit. Olive baboons at Gombe are known to consume V. doniana along with several other Vitex sp. (e.g., V. cumini, V. tharagunus, V. madiensis,

V. fischer) (Collins, pers. comm.). Thus, PdG would likely not have been a reliable reproductive hormone to use to assess ovulatory functioning in this population even if

RIA were possible.

Furthermore, because female olive baboons mate with fewer partners during estrous cycles than other polygynandrous species, this likely also influences their copulation rates. A relatively small sample of copulations especially limited my ability to test the impact of copulation context on the expression of female post-copulatory signals.

In particular, I was only able to record a limited sample of copulation calls. The quiet nature of copulation calls in olive baboons also made it difficult to record many high- quality calls that could be assessed for both temporal features (duration, number of units, etc.) and other acoustic features, like fundamental frequencies, that could potentially be

154 compared to recordings from other species, especially other species of savanna baboon

(e.g., yellow baboons: Semple 2001). Thus, an important research endeavor in the future, along with conducting playback experiments to assess the referential nature of calls (see

Future Directions below) is to record more high-quality calls from olive baboons at

Gombe and other research sites. The use of new technology, like radio collars that can simultaneously record spatial positioning and vocalizations are being experimented with in field settings (O’Brian, pers. comm.) and are an exciting prospect for recording particularly quiet calls like olive baboon copulation calls in the future.

It was somewhat surprising that current male rank was not a greater predictor of either female preferences or male copulatory success in a variety of contexts. It is likely that lower ranking males were less represented in the sample simply because male competition limits their ability to interact with females. This is especially true of

“peripheral” males, or those just beginning to integrate into the group. Over time, more observations would likely warrant important distinctions between female copulations with males who are fully integrated and those who are less so, and may be classified as lower ranking as a result, and this will be an important endeavor for future studies of female preferences in this population and others.

Future Directions

Along with addressing the above limitations in my study, there are several important research agendas that have been born out of my current project. Even in studies like mine that are careful to consider the potentially confounding or interacting effects of male aggression and coercion and female-female competition on female mate preferences

155 and also consider multiple male qualities that may influence these preferences, it is important to additionally consider longitudinal information that can demonstrate the adaptive significance of female choice. To demonstrate that female mate choice in primates is an evolutionary force, rather than a “behavioral phenomenon" (Keddy-Hector

1992) it needs to be demonstrated that females derive significant reproductive benefits from both pre-copulatory and post-copulatory mechanisms of female mate choice and whether these mechanisms impact offspring fitness (Paul 2002; Setchell & Kappeler

2003; Shanoor & Jones 2003).

To test the hypotheses developed in this dissertation, it is especially necessary to determine whether males who females mate with frequently during fertile periods, especially during conceptive cycles, contribute to the care of infants in ways that can benefit offspring through effects like accelerated maturation. This relationship has been demonstrated in yellow baboons, in which it has also been demonstrated through DNA analysis that males sire offspring who benefit most from these reproductive benefits

(Buchan et al. 2003; Charpentier et al. 2008). Given the similarities between olive baboons and yellow baboons, we might expect similar results in olives. These types of analyses requires longitudinal studies that can (a) demonstrate robustness of relationships between females and their preferred partners to determine whether they are maintained as friendships during the vulnerable stages of infant development (b) incorporate appropriate genetic evidence that can determine relatedness between males and offspring

(c) determine if males do not sire offspring, but do provide care, whether there are additional benefits for males in these relationships (e.g., through increased mating access

156 with females) (d) determine the potential long-term benefits for offspring of having male protectors in the group (e.g., acceleration of maturation as in yellow baboons)

(Charpentier et al. 2008). This is the primary future research agenda for individuals involved in the Gombe Stream olive baboon research project and requires data from multiple generations and appropriate training in the collection and processing of fecal samples that can be assessed for appropriate DNA evidence. Given the opportunity, these types of data will be significant for continuing the efforts of my study, comparative analyses across savanna baboon species, and the broader efforts of primatologists attempting to address the intersecting effects of mechanisms of sexual selection.

Given that it appears male social inertia, or the ability to remain in a troop for extended time periods, may be an important mechanism either driving female preferences or influencing male-male competition in the Gombe population, future studies should consider whether this social stability is assessable to females in ways other than male characteristics like current age, tenure, rank, and rank trajectory. I also hope to address, through collaborations with individuals studying chimpanzee ranging and hunting patterns at Gombe, the significance of chimpanzee predation on Gombe baboon populations and test more directly whether chimpanzee hunting is a selective pressure driving female preferences for protective males in the Gombe baboons.

Additionally, at Gombe, it may be possible for some males to remain in their groups for more protracted periods of time due to their relatively stable food sources and unique forested habitat (Packer et al. 1995). Individuals in DC troop (commonly referred to as “Camp Troop”) may be benefiting from food stability because their home range

157 includes the main residential area for Gombe researchers, tourists, and Tanzania National

Parks employees and they supplement their diets with human food scraps. Thus, this effect of social inertia, or social stability, may be more influential on female preferences in the Gombe population, or similar forest-dwelling baboons or groups that are able to derive some resource benefits from overlapping ranges with humans and thus there may be greater opportunities for certain males to have extended tenures. Comparative studies within olive baboons and across species can help determine whether the effect of male social stability is group, population, or species specific or whether this is a quality that characterizes female mate preferences more generally, particularly in multimale, multifemale primate societies.

Furthermore, more laboratory and experimental approaches with captive and wild populations are needed to test alternative hypotheses concerning different selective pressures (e.g. direct vs. indirect benefits) on the evolution of female mate preferences among primates (Setchell & Kappeler 2003). These studies can be difficult to pursue, in part because there are more heavy restrictions posed on the manipulation of primates than other animals. However, more studies are being conducted on captive populations and approaches that use noninvasive techniques have been incorporated into wild and free- ranging primate studies (Soltis et al. 1999b; Soltis et al. 2001; Paul 2002; Nikitopoulos et al. 2005). For example using a pair-choice test in captivity that allowed females control over their interactions with males, Nikitopoulos and colleagues (2005) demonstrated female long-tailed macaques copulate at equal rates with all available partners across their ovulatory cycles. Females did not discriminate amongst males even on the basis of

158 male dominance rank, suggesting promiscuity may be the “preferred” pattern in this species (Nikitopoulos et al. 2005). Studies like these conducted on a broader range of species will be an important research avenue for primatologists testing female choice using experimental designs that control for confounding effects of male competition.

Experimentation is also one of the most important avenues for further assessing the function of copulation calls in olive baboons as well as other catarrhine primates with diverse social structures and vary in the ways and degrees in which male and female reproductive strategies interact (e.g., relative significance of infanticide as a male coercive strategy). Playback experiments should be used in future tests of the referential signaling of copulation calls (Macedonia & Evans 1993; Engelhardt et al. 2012) to determine whether copulation calls change the behavior of the current partner as well as other listening males. This is an especially important endeavor in investigating the different functions of copulation calls in the Papio genus. Playback experiments should be carefully replicated in chacma, yellow, and olive baboons to determine whether the reactions of males to female calling are more similar or different based on the relative influence of female choice and male competition/coercion in these different species.

Additionally, in the future, I will directly compare copulation calls that have been recorded from a wild population of chacma baboons in the Moremi Game Reserve,

Botswana (Walz & Kitchen, unpublished data) with my recordings from the Gombe olive baboons to systematically determine how these calls differ in their average number of units, unit rates, exhale rates, etc., and the contexts in which these calls are produced.

159

Finally, evidence from my study as well as olive baboons at Laikipia, Kenya

(Danish & Palombit 2014) indicates that consort contest behavior in olive baboons is a complex alternative reproductive strategy (similar to followers and satellites in birds, amphibians and fish: Gross 1996) that can allow males of all ranks to successfully copulate and consort with females. These results suggest that the role of male contests in future studies of the interacting strategies of males and females should focus on male followers beyond just the descriptive reports of this alternative reproductive strategy that have been previously produced (reviewed by Danish & Palombit 2014). In particular, given that my study suggests female preferences may drive male success in consorting with females after investing time in a contest, these interactions deserve greater consideration in the future. It will also be important to assess other male-female social and post-copulatory behaviors and interactions, and I next plan to use the non-copulatory social behavior that I collected to examine dissociation and solicitation rates as well as disruption patterns in the Gombe groups. Additionally, my study demonstrated that there are possible important cost-benefit trade-offs for females engaging in consorts. They may allow females to mate more consistently with preferred partners, but also limit ability to mate with others. Engaging in consorts also seems to provide female olive baboons some protection from aggression from other females. Moving forward, I hope to address through analysis of fecal glucocorticoid concentrations (Sapolsky 1983), whether the consort context limits or contributes to stress females experience throughout their reproductive lives and whether this varies according to female rank and reproductive experience.

160

Through this research I have contributed to the long-standing tradition in primatology of examining the interactions and complexities in mechanisms of sexual selection acting on the lives of female baboons, and demonstrate the costs and benefits females encounter as they navigate their social and reproductive lives. Furthermore, in many ways, early hominin social systems are convergent with those expressed in savanna baboons, making them important models for testing female strategies within complex social groups in which competition for mates and resources dominates (Smuts 1985).

Addressing female mate preferences from multiple levels, as I have done here, can provide a comprehensive picture of the strategies female primates within the context of this social milieu can use to gain male support and form social bonds, and could also reflect the specific selective that influenced the evolution of male-female bonds among humans. Future research endeavors will continue to examine potential fitness consequences for females who bias paternity towards a few preferred partners using both pre- and post-copulatory sexual signals, tease apart what might be the characteristics females are choosing, and assess whether competitive females also derive significant reproductive benefits from directing aggression towards competitors.

161

REFERENCES

Alberts SC, Watts HE, Altmann J (2003) Queuing and queue-jumping: long-term patterns of reproductive skew in male savannah baboons, Papio cynocephalus. Anim Behav 65:821-840

Altmann J (1980) Baboons mothers and infants. Harvard University Press, Cambridge

Altmann J, Alberts SC (2003) Variability in reproductive success viewed from a life- history perspective in baboons. Am J Hum Biol 15:401-409

Altmann SA (1962) A field study of the sociobiology of rhesus monkeys, Macaca mulatta. Ann NY Acad Sci 102:338-435

Altmann SA (1973) The pregnancy sign in savannah baboons. Jour Zoo Anim Med 4:8- 12

Altmann SA (1974) Baboons, space, time, and energy. Amer Zool 14:221-248

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

Aujard F, Heistermann M, Theirry B, Hodges JK (1998) Functional significance of behavioral, morphological, and endocrine correlates across the ovarian cycle in semifree ranging Tonkean macaques. Am J Primatol 46:285-309

Bailey A, Eberly LE, Packer P (2015) Does pregnancy coloration reduce female conspecific aggression in the presence of maternal kin? Anim Behav 108:199-206

Barton RA, Whiten A (1993) Feeding competition among female olive baboons Papio anubis. Anim Behav 46:777-789

Bateman AJ (1948) Intra-sexual selection in Drosophila. Heredity 2:349-368.

Beach FA (1976) Sexual attractivity, proceptivity, and receptivity in female mammals. Horm Behav 7:105-138

Beehner JC, Whitten PL (2004) Modification of a field method for fecal steroid analysis in baboons. Phys Behav 82:269-277

162

Beehner JC, Bergman TJ, Cheney DL, Seyfarth RM, Whitten PL (2006) Testosterone predicts future dominance rank and mating activity among chacma baboons. Behav Ecol Sociobiol 59:469-479

Behboodi E, Katz DF, Samuels SJ, Tell L, Hendrickx AG, Lasley BL (1991) The use of a urinary estrone conjugates assay for detection of optimal mating time in the cynomolgus macaque (Macaca fascicularis). J Med Primatol 20:229-234

Bergman TJ, Beehner JC, Cheney DL, Seyfarth RM, Whitten PL (2005) Correlates of stress in free-ranging male chacma baboons, Papio hamadryas ursinus. Anim Behav 70:703-713

Bercovitch FB (1985) Reproductive tactics in adult female and adult male olive baboons. PhD Dissertation, University of California, Los Angeles

Bercovitch FB (1987) Reproductive success in male savanna baboons. Behav Ecol Sociobiol 21(3):163-72

Bercovitch FB (1988) Coalitions, cooperation and reproductive tactics among adult male baboons. Anim Behav 36:1198-1209

Bercovitch FB (1989) Body size, sperm competition, and determinants of reproductive success in male savanna baboons. Evolution 43:1507-21

Bercovitch FB (1991) Mate selection, consortship formation, and reproductive tactics in adult female savanna baboons. Primates 32:437-52

Bercovitch FB (1995) Female cooperation, consortship maintenance, and male mating success in savanna baboons. Anim Behav 50:137-149

Bercovitch FB, Strum SC (1993) Dominance rank, resources availability, and reproductive maturation in female savanna baboons. Behav Ecol Sociobiol 33:313-318

Berenstain L, Wade TD (1983) Intrasexual selection and male mating strategies in baboons and macaques. Int J Primatol 4:201-236.

Birkhead TR, Pizzari T (2002) Postcopulatory sexual selection. Nat Rev Genet 3:262-273

Blaustein JD (2009) Feminine reproductive behavior and physiology in rodents: integration of hormonal, behavioral, and environmental influences. In: Pfaff DW, Arnold PA, Etgen AM, Fahrbach SE, and Rubin RT (eds) Hormones, brain and behavior. Elsevier, San Diego, pp 67-107

163

Boese GK (1975) Social behavior and ecological considerations of West African baboons. In: Tuttle RH (ed) Socioecology and psychology of primates. Mouton, The Hague, pp 205-230

Buchan JC, Alberts SC, Silk JB, Altman J (2003) True paternal care in a multi-male primate society. Nature 425:179-181

Bulger JB (1993) Dominance rank and access to estrous females in male savanna baboons. Behav 127:67-10.

Burnham KP, Anderson DR, Huyvaert KP (2011) AIC model selection and multimodel inference in behavioral ecology: some background, observations, and comparisons. Behav Ecol Sociobiol 65:23-35

Busse C, Hamilton WJ (1981) Infant carrying by male chacma baboons. Science 212: 1281-1283.

Chapman T, Arnqvist G, Bangham J, Rowe L (2003) Sexual conflict. Trends Ecol Evol 18:41-47.

Charnov EL, Berrigan D (1993) Why do female primates have such long lifespans and so few babies? Or life in the slow lane. Evol Anthropol 6:191-194.

Charpentier MJE, Van Horn RC, Altmann J, Alberts SC (2008) Paternal effects on offspring fitness in a multimale primate society. Proc Natl Acad Sci USA 105:1988-1992

Cheney DL, Seyfarth RM (2007) Baboon metaphysics. University of Chicago Press, Chicago

Cheney DL, Moscovice LR, Heesen M, Mundry R, Seyfarth RM (2010) Contingent cooperation between wild female baboons. Proc Natl Acad Sci USA 17:9562-9566

Cheney DL, Silk JB, Seyfarth RM (2012) Evidence for intrasexual selection in wild female baboons. Anim Behav 84:21-27

Cheng MF (1992) For whom does the female dove coo? A case for the role of vocal self- stimulation. Anim Behav 43:1035-1044

Clarke PMR, Pradhan GR, van Schaik CP (2009) Intersexual conflict in primates: Infanticide, paternity allocation, and the role of coercion. In: Muller MN, Wrangham RW (eds) Sexual coercion in primates and humans. Harvard University Press, Cambridge, pp 42–77

Clutton-Brock TH (2009) Sexual selection in females. Anim Behav 77:3-11

164

Clutton-Brock TH, Gillett JB (1979) A survey of forest composition in the Gombe Stream National Park, Tanzania. Afr J Ecol 17:131-158

Clutton-Brock TH, Parker GA (1995) Sexual coercion in animal societies. Animal Behaviour, 49:1345–1365.

Collins DA (1981) Behaviour and patterns of mating among adult yellow baboons (Papio cynocephalus) Ph.D. dissertation. University of Edinburgh, Edinburgh

Cox CR, Leboeuf BJ (1977) Female incitation of male competition: A mechanism in sexual selection. Amer Nat 111:317-335

Crockford C, Wittig RM, Seyfarth RA, Cheney DL (2007) Baboons eavesdrop to deduce mating opportunities. Anim Behav 73:885-890.

Crockett CM, Sekulic R (1984) Infanticide in red howler monkeys (Alouatta seniculus). In: Hausfater G, Hrdy SB (eds) Infanticide: comparative and evolutionary perspectives. Aldine Publishing Company, New York, pp 173-191

Danish LM, Palombit RA (2014) “Following,” an alternative mating strategy used by male olive baboons (Papio hamadyras anubis): quantitative behavioral and functional description. Int J Primatol 35:394-410.

Darwin C (1871) The descent of man and selection in relation to sex. John Murray, London

Daspre A, Heistermann M, Hodges JK, Lee PC, Rosetta L (2009) Signals of female reproductive quality and fertility in colony-living baboons (Papio h. anubis) in relation to ensuring paternal investment. Am J Primatol 71:529-538

De Ruiter JR, van Hooff JARAM, Scheffrahn W (1994) Social and genetic aspects of paternity in wild long-tailed macaques (Macaca fascicularis). Behaviour 129:203-224

Deschner T, Heistermann M, Hodges K, Boesch C (2004) Female sexual swelling size, timing of ovulation, and male behavior in wild West African chimpanzees. Horm Behav 46(2):204-215

Dixson A (1998) Primate sexuality: comparative studies of the prosimians, monkeys, apes, and human beings. Oxford University Press, Oxford

Domb LG, Pagel M (2001) Sexual swellings advertise female quality in wild baboons. Nature 410:204-206

165

Dunbar RIM (1984) Reproductive decisions: an economic analysis of gelada baboon social strategies. Princeton University Press, Princeton

Eisenfeld AJ (1969) Hypothalamic estradiol binding macromolecules. Nature 224:1202- 1203

Elo AE (1978) The rating of chess players, past and present. Arco, New York

Engelhardt A, Fischer J, Neumann C, Pfeifer JB, Heistermann M (2012) Information content of female copulation calls in wild long-tailed macaques (Macaca fascicularis). Behav Ecol Sociobiol 66:121-134

Engelhardt A, Heistermann M, Hodges JK, Nurnberg P, Niemitz C (2006) Determinants of male reproductive success in wild-long-tailed macaques (Macaca fascicularis)—Male monopolisation, female mate choice or post-copulatory mechanisms. Behav Ecol Sociobiol 59:740–752

Erhart EM, Overdorff DJ (1998) Infanticide in Propithecus diadema edwardsi: an evaluation of the sexual selection hypothesis. Intl J Primatol 19:73-81

Fallon BL, Neumann C, Byrne RW, Zuberbuhler K (2016) Female chimpanzees adjust copulation calls according to reproductive status and level of female competition. Anim Behav 113:87-92

Fedigan LM (1983) Dominance and reproductive success in primates. Yrb Phys Anth 26:91-129

Fischer J, Kitchen DM, Seyfarth RM, Cheney DL (2004) Baboon loud calls advertise male quality: Acoustic features and their relation to rank, age, and exhaustion. Behav Ecol Sociobiol 56:140–148.

Fisher RA (1915) The evolution of sexual preference. Eugenics Review 7:184-192

Floody OR (1983) Hormones and aggression in female mammals. In: Svare BB (ed) Hormones and aggressive behavior. Plenum Press, New York, pp 39–89

Gaines SD, Rice WR (1990) Analysis of biological data when there are ordered expectations. Am Nat 135:310– 317.

Gesquiere LR, Wango EO, Alberts SC, Altmann J (2007) Mechanisms of sexual selection: sexual swellings and estrogen concentrations as fertility indicators and cues for male consort decisions in wild baboons. Horm Behav 51:114-125.

Girard-Buttoz C, Heistermann M, Rahmi E, Agil M, Fauzan PA , Engelhardt A (2014)

166

Costs of and investment in mate-guarding in wild long-tailed macaques (Macaca fascicularis): influences of female characteristics and male-female social bonds. Int J Primatol 35:701-724.

Gouzoules H, Gust DA, Donaghey B, St. Andre E (1998) Estrus vocalizations in two primate species (Cercocebus torquatus atys and Macaca nemestrina): evidence for an effect of intrasexual competition. Evol Comm 2:189–215

Gowaty PA, Kim YK, Anderson WW (2012) No evidence of sexual selection in a repetition of Bateman’s classic study of Drosophila melanogaster. Proc Natl Acad Sci USA 109:11740–11745

Gowaty PA, Hubbell SP (2009) Reproductive decisions under ecological constraints: it's about time. Proc Natl Acad Sci USA 106:10017-10024

Gross MR (1996) Alternative reproductive strategies and tactics: diversity within the sexes. Trends Ecol Evol 11: 92-98.

Hall KRL, DeVore I (1965) Baboon social behavior. In: DeVore I (ed.), Primate behavior: Field studies of monkeys and apes. Holt, Rinehart, and Winston, New York, pp 53-110

Hamilton WD (1964) The genetical evolution of social behaviour. I and II. J Theor Bio 7:1-52.

Hamilton WJ, Arrowood PC (1978) Copulatory vocalizations of chacma baboons (Papio usinus), gibbons (Hylobates lar), and humans (Homo sapiens). Science 200:1405-1409

Harcourt AH (1997) Sperm competition in primates. Amer Nat 149:189-194

Hausfater G (1975) Dominance and reproduction in baboons (Papio cynocephalus): A quantitative analysis. Contrib Primatol 7:2-150

Heistermann M, Ziegler T, Van Schaik C P, Launhardt K, Winkler P, Hodges JK (2001) Loss of oestrus, concealed ovulation and paternity confusion in free-ranging Hanuman langurs. Proc R Soc Lond B 268(1484):2445-2451

Henzi SP, Clarke PMR, van Schaik CP, Pradhan GR, Barrett L (2010) Infanticide and reproductive restraint in a polygynous social mammal. Proc Natl Acad Sci 107:2130- 2135

Henzi SP, Barrett L (2005) The historical socioecology of savanna baboons (Papio hamadryas). J Zool 265:215–226

167

Higham JP, Ross C, Warren Y, Heistermann M, MacLarnon AM (2007) Reduced reproductive function in wild baboons (Papio hamadryas anubis) related to natural consumption of the African black plum (Vitex doniana). Horm Behav 52:384-390

Higham JP, Heistermann M, Ross C, Semple S, MacLarnon A (2008) The timing of ovulation with respect to sexual swelling detumescence in wild olive baboons. Primates 4:295-299

Hohmann G, Fruth B (2003) Intra- and inter-sexual aggression by bonobos in the context of mating. Behaviour 140:1389–1413

Hrdy SB (1999) The woman that never evolved. Harvard University Press, Cambridge

Huchard E, Courtiol A, Benavides JA, Knapp LA, Raymond M, Cowlishaw G (2009) Can fertility signals lead to quality signals? Insights from the evolution of primate sexual swellings. Proc R Soc Lond B 276:1889-1897

Huchard E, Cowlishaw G (2011) Female-female aggression around mating: an extra cost of sociality in a multimale primate society. Behav Ecol 22:1003-1011.

Inoue E, Takenaka O (2008) The effect of male tenure and female mate choice on paternity in free-ranging Japanese macaques. Am J Primatol 70:62-68

Jeffcoate SL (1983) Use of rapid hormone assays in the prediction of ovulation. In: Jeffcoate SL (ed) Ovulation: methods for its prediction and detection. John and Wiley Sons Ltd, Chichester, pp. 67–82

Johnson JA (1984) Social relationships of juvenile olive baboons. PhD Thesis, University of Edinburgh, Edinburgh

Johnson Elizabeth T, Snyder-Mackler N, Beehner JC, Bergman TJ (2014) Kinship and dominance rank influence the strength of social bonds in female geladas (Theropithecus gelada). Int J Primatol 35:288-304

Jordan WC, Bruford MW (1998) New perspectives on mate choice and the MHC. Heredity 81:127-133

Kaburu SSK, Newton-Fisher NE (2015) Trading or coercion? Variation in male mating strategies between two communities of East African chimpanzees. Behav Ecol Sociobiol 69:1039-1052.

Kapsalis E (2003) Matrilinieal kinship and primate behavior. In: Chapais B, Berman C (eds) Kinship and behavior in primates. Oxford University Press, Oxford, pp 153-176

168

Kappeler PM (2012) Mate choice. In: Mitani JC, Call J, Kappeler PM, Palombit RP, Silk JB (eds) The evolution of primate societies. The University of Chicago Press, Chicago, pp 367-386

Kappeler PM, van Schaik CP (2002) Evolution of primate social systems. Int J Primatol 23:707-740

Keddy-Hector AC (1992) Mate choice in non-human primates. Amer Zool 32:62-70

Kitchen DM, Seyfarth RM, Fischer J, Cheney DL (2003) Loud calls as indicators of dominance in male baboons (Papio cynocephalus ursinus) Behav Ecol Sociobiol 53:374- 384

Kitchen DM, Cheney DL, Seyfarth RM (2005) Male chacma baboons (Papio hamadryas ursinus) discriminate loud call contests between rivals of different relative ranks. Anim Cogn 8:1-6.

Kitchen DM, Beehner JC, Bergman TJ, Cheney DL, Crockford C, Engh AL, Wittig R (2009a) The causes and consequences of male aggression directed at female chacma baboons. In: Muller MN, Wrangham RW (eds) Sexual coercion in primates and humans. Harvard University Press, Cambridge, pp 128-156

Kitchen DM, Cheney DL, Seyfarth RM, Beehner JC (2009b) A link between fecal testosterone and an honest signal – the loud ‘wahoo’ vocalizations of chacma baboons. Am J Phys Anthropol S48:166

Krebs J, Davies NB (1993) Sexual conflict and sexual selection: an introduction to behavioural ecology. 3rd ed. Blackwell Scientific Publications, London

Lemasson A, Palombit RA, Jubin R (2008) Friendships between males and lactating from females in a free-ranging group of olive baboons (Papio hamadryas anubis): evidence playback experiments. Behav Ecol Sociobiol 62:1027-1035

Lombardi CM, Hurlbert SH (2009) Misprescription and misuse of one-tailed tests. Aus Ecol 34:447-468.

Loy J (1971) Estrous behavior of free-ranging rhesus monkeys (Macaca mulatta). Primates 12:1-31

Loy J, Loy K (1977) Sexual harassment among captive patas monkeys (Erythrocebus patas). Primates 18:691-699.

169

Macedonia JM, Evans CS (1993) Variation among mammalian alarm call systems and the problem of meaning in animal signals. Ethology 93:177-197

Machanda ZP, Gilby IC, Wrangham RW (2013) Male-female association patterns among free-ranging chimpanzees (Pan troglodytes schweinfurthii) Int J Primatol 34:917-938

Maestripieri D, Roney JR (2005) Primate copulation calls and post-copulatory female choice. Behav Ecol 16:106-113.

Manson JH (1992) Measuring female mate choice in Cayo Santiago rhesus macaques. Anim Behav 44:405-416.

Matsumoto-Oda A (1999) Female choice in the opportunistic mating of wild chimpanzees. (Pan troglodytes schweinfurthii) at Mahale. Behav Ecol Sociobiol 58:247- 255.

Melnick D, Pearl M (1986) Cercopithecines in multi-male groups: genetic diversity and population structure. In: Cheney DL, Smuts BB, Seyfarth RM, Wrangham RW, Struhsaker TT (eds) Primate societies. University of Chicago Press, Chicago, pp 121-134

Møller AP, Birkhead TR (1989) Copulation in mammals: evidence that sperm competition is widespread. Biol J Linn Soc 38:119-131

Moscovice LR, Heesen M, Di Fiore A, Seyfarth RM, Cheney DL (2009) Paternity alone does not predict long-term investment in juveniles by male baboons. Behav Ecol Sociobiol 63:1471-1482

Moscovice LR, Di Fiore A, Crockford C, Kitchen DM, Wittig R, Seyfarth RM, Cheney DL (2010) Hedging their bets? Male and female chacma baboons form friendships based on likelihood of paternity. Anim Behav 79:1007-1015

Muller MN; Thompson ME; Kahlenberg SM, Wrangham RW (2011) Sexual coercion by male chimpanzees shows that female choice may be more apparent than real. Behav Ecol Sociobiol 65(5):921-933

Muller MN, Wrangham RW (2009) Sexual coercion in primates and humans: an evolutionary perspective on male aggression against females. Harvard University Press, Cambridge

Neumann C, Duboscq J, Dubuc C, Ginting A, Maulana I, Agil M, Widdig A, Engelhardt A (2011) Assessing dominance hierarchies: validation and advantages of progressive evaluation with Elo-rating. Anim Behav 82:911-921

Neumann C, Kulik L (2014) Animal dominance hierarchies by Elo-rating. R Package

170

‘EloRating’

Niemeyer CL, Anderson JR (1983) Primate harassment of matings. Ethol Sociobiol 4:205-220

Nikitopoulos E, Arnhem E, van Hooff J, Sterck E (2004) Influence of female copulation calls on male sexual behavior in captive Macaca fascicularis. Int J Primatol 25: 659-677

Nikitopoulos E, Heistermann M, de Vries H, van Hooff JARAM, Sterck EHM (2005) A pair choice test to identify female mating pattern relative to ovulation in longtailed macaques, Macaca fascicularis. Anim Behav 70:1283-1296

Nunn CL (1999) The number of males in primate social groups: a comparative test of the socioecological model. Behav Ecol Sociobiol 46:1-13

O’Connell SM, Cowlishaw G (1994) Infanticide avoidance, sperm competition and mate choice -the function of copulation calls in female baboons. Anim Behav 44:687-694.

O’Connell SM, Cowlishaw G (1995) The post-copulation withdrawal response in female baboons - a functional-analysis. Primates 36:441-446

Oda R, Masataka N (1992) Functional significance of female Japanese macaque copulatory calls. Folia Primatol 58:146-149.

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

Packer C (1979b) Male dominance and reproductive activity in Papio anubis. Anim Behav 27: 37-46

Packer C (1980) Male care and exploitation of infants in Papio anubis. Anim Behav 28: 512-20

Packer C, Pusey AE (1979) Female aggression and male membership in troops of Japanese macaques and olive baboons. Folia Primatol 31:212-18

Packer C, Collins DA, Sindimwo A, Goodall J (1995) Reproductive constraints on aggressive competition in female baboons. Nature 373:60-63

Palombit RA, Seyfarth RM, Cheney DL (1997) The adaptive value of 'friendships' to female baboons: experimental and observational evidence. Anim Behav 54:599-614

Palombit RA (2000) Infanticide and the evolution of male-female bonds in animals. In: van Schaik CP, Janson CH (ed) Infanticide by males and its implications. Cambridge

171

University Press, Cambridge, pp 239-268

Palombit RA, Cheney DL, Fischer J, Johnson S, Rendall D, Seyfarth RM Silk JB (2000) Male infanticide and defense of infants in chacma baboons. In: van Schaik CP, Janson CH (eds) Infanticide by males and its implications. Cambridge University Press, Cambridge, pp 123-152

Palombit RA, Cheney DL, Seyfarth RM (2001) Female-female competition for male 'friends' in wild chacma baboons, Papio cynocephalus ursinus. Anim Behav 61:1159 – 1171

Palombit RA (2009) "Friendships" with males: A female counterstrategy to infanticide in chacma baboons of the Okavango Delta. In: Muller MN, Wrangham RW (ed) Sexual coercion in primates and humans: an evolutionary perspective on male aggression against females. Harvard University Press, Cambridge, pp 377-409

Palombit RA (2014) Sexual conflict in nonhuman primates. Adv St Behav 46:191-280

Paul A (2002) Sexual selection and mate choice. Int J Primatol 23:877-904

Pfaus JG (1999) Neurobiology of sexual behavior. Curr Op Neurobiol 9:751-758

Pfefferle D, Brauch K, Heistermann M, Hodges JK, Fischer J (2008) Female barbary macaques (Macaca sylvanus) copulation calls do not reveal the fertile phase but influence mating outcomes. Proc R Soc Lond B 275:571-578

Pradhan GR, Engelhardt, A, van Schaik CP, Maestripieri D (2006) The evolution of female copulation calls in primates: A review and new model. Behav Ecol Sociobiol 59:333-343

Pusey A (2012) Magnitude and sources of variation in female reproductive performance. In: Mitani JC, Call J, Kappeler PM, Palombit RP, Silk JB (eds) The evolution of primate societies. The University of Chicago Press, Chicago, pp 343-366

Puts DA, Dawood K (2006) The evolution of female orgasm: adaptation or byproduct? Twin Res Hum Genet 9:467-472

Rakhovskaya MV (2013) Correlates of male consortship rate in free-ranging rhesus macaques (Macaca mulatta) Int J Primatol 34:662-680

Ransom TW (1981) Beach troop of the Gombe. Bucknell University Press, East Brunswick

172

Ransom TW, Ransom BS (1971) Adult male-infant relations among baboons (Papio anubis). Folia Primatol 16:179-195

Rasmussen KLR (1980) Consort behavior and mate selection in yellow baboons (Papio cynocephalus) PhD Thesis, University of Cambridge, Cambridge

Rigaill L, Higham JP, Lee PC, Blin A, Garcia C (2013) Multimodal sexual signaling and mating behavior in olive baboons (Papio anubis). Am J Primatol 75:774-787

Saayman GS (1970) The menstrual cycle and sexual behavior in a troop of free-living chacma baboons (Papio ursinus). Folia Primatol 12:81-110

Sapolsky RM (1983) Endocrine aspects of social instability in the olive baboon (Papio anubis). Am J Primatol 5:365-379

Scelza BA (2013) Choosy but not chaste: multiple mating in human females. Evol Anthropol 22:259-269

Semple S (1998) The function of Barbary macaque copulation calls. Proc R Soc Lond B 265:287-291

Semple S (2001) Individuality and male discrimination of female copulation calls in the yellow baboon. Anim Behav 61:1023-1028

Semple S, McComb K, Alberts S, Altmann J (2002) Information content of female copulation calls in yellow baboons. Am J Primatol 56:43-56.

Semple S (2001) Individuality and male discrimination of female copulation calls in the yellow baboon. Anim Behav 61:1023-1028

Setchell JM, Kappeler PM (2003) Selection in relation in sex in primates Adv St Behav 33:87-173

Setchell JM, Charpentier MJ, Abbott KM, Wickings EJ, Knapp LA (2010) Opposites attract: MHC-associated mate choice in a polygynous primate. J Evol Biol 23:136-148

Seyfarth RM (1977) A model of social grooming among adult female monkeys. J Theor Biol 65:671-698

Seyfarth RM (1978) Social relationships among adult male and female baboons, I. Behaviour during sexual consortship. Behaviour 64:204-226

Shaikh A, Celaya C, Gomez I, Shaikh S (1982) Temporal relationship of hormonal peaks to ovulation and sex skin deturgescence in the baboon. Primates 22:444-452

173

Shanoor N, Jones CB (2003) A brief history of the study of sexual selection and reproductive competition in primatology. In: Jones CB (ed) Sexual selection and reproductive competition in primates: new perspectives and directions. American Society of Primatologists, Norman, pp 1-43

Shur M (2008) The socioendocrinology of “friendship” between adult male and lactating female baboons. Ph.D. Thesis, Rutgers University, New Jersey.

Silk JB, Altmann J, Alberts SC (2006) Social relationships among adult female baboons (Papio cynocephalus) I. Variation in the strength of social bonds. Behav Ecol Sociobiol 61:183-195.

Silk JB, Beehner JC, Bergman TJ, Crockford C, Engh AL, Moscovice LR, Wittig RM, Seyfarth RM, Cheney DL (2010) Female chacma baboons form strong, equitable, and enduring social bonds. Behav Ecol Sociobiol 64:1733-1747

Small MF (1984) Female primates: studies by women primatologists. Alan R. Liss, New York

Small MF (1988) Female primate sexual behaviour and conception. Are there really sperm to spare? Curr Anthropol 29:81-100.

Small MF (1989) Female choice in nonhuman primates. Ybk Phys Anthropol 32:103- 127.

Small M.J. (1993). Female choices. Cornell University Press, Cornell

Smith DG (1994) Male-dominance and reproductive success in a captive group of rhesus macaques (Macaca mulatta). Behavior 129:225-242

Smuts BB (1985) Sex and friendship in baboons. Harvard University Press, Cambridge

Smuts BB, Smuts RW (1993) Male aggression and sexual coercion of females in nonhuman-primates and other mammals - evidence and theoretical implications. Adv St Behav 22:1-63

Soltis J, Mitsunaga F, Shimizu K, Yanagihara Y, Nozaki M (1997) Sexual selection in Japanese macaques I: female mate choice or male sexual coercion? Anim Behav 54:725– 736

Soltis J (1999a) Measuring male-female relationships during the mating season in wild Japanese macaques (Macaca fuscata yakui). Primates 40:453-467

174

Soltis J (1999b) Female mating strategy in an enclosed group of Japanese macaques. Am J Primatol 47:263-278.

Soltis J, Thomsen R, Matsubayashi K, Takenaka O (2000) Infanticide by resident males and female counter-strategies in wild Japanese macaques (Macaca fuscata). Behav Ecol Sociobiol 48:195-202

Soltis J, Thomsen R, Takenaka O (2001) The interaction of male and female reproductive strategies and paternity in wild Japanese macaques, Macaca fuscata. Anim Behav 62:485-494

Sommer V (1989) Sexual harassment in langur monkeys (Presbytis entellus): competition for ova, sperm, and nurture. Ethol 80:205-217

Strum SC (1975) Life with the Pumphouse Gang: New insights into baboon behavior. Nat Geogr147:672-691

Strum SC (1982) Agonistic dominance in male baboons: An alternative view Inter J Primatol 3:175-202

Strum SS, Latour B (1987) Redefining the social link - from baboons to humans. Social Science Information Sur Les Sciences Sociales 26:783-802

Stumpf RM, 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 RM, Boesch C (2006) The efficacy of female choice in chimpanzees of the Taï National Park, Côte d'Ivoire. Behav Ecol Sociobiol 60:749-765

Stumpf RM, Thompson ME, Knott CD (2008) A comparison of female mating strategies in Pan troglodytes and Pongo spp. Int J Primatol 29:865-884

Swedell L, Schreier A (2009) Male aggression towards females in Hamadryas baboons: conditioning, coercion, and control. In: Muller M, Wrangham R (ed) Sexual coercion in primates: an evolutionary perspective on male aggression against females. Harvard University Press, Cambridge, pp 244-268

Tinbergen N (1963) On aims and methods of ethology. Zeitschrift für Tierpsychologie 20:410-433.

Todt D, Hammerschmidt K, Ansorge V, Fischer J (1995) The vocal behavior of Barbary macaques (Macaca sylvanus): call features and their performance in infants and adults In: Zimmermann E, Newman J, Juergens U (eds). Current topics in primate vocal communications. Plenum Press, New York, pp 141-160 175

Townsend SW, Deschner T, Zuberbühler K (2008) Female chimpanzees use copulation calls flexibly to prevent social competition. PloS One 3:e2431

Townsend SW, Deschner T, Zuberbühler, K (2011) Copulation calls in female chimpanzees (Pan troglodytes schweinfurthii) convey identity but do not accurately reflect fertility. Int J Primatol 32:914-923

Trivers RL (1972) Parental Investment and Sexual Selection. In: Campbell B (ed) Sexual selection and the descent of man (1871-1971). Aldine Publishing Company, Chicago, pp 136-179 van Noordwijk MA, van Schaik CP (1987) Competition among female long-tailed macaques, Macaca fascicularis. Anim Behav 35:577-589 van Schaik CP, van Noordwijk MA, Nunn CL (1999) Sex and social evolution in primates. In: Lee PC (ed) Comparative primate socioecology. Cambridge University Press, Cambridge, pp 204-240 van Schaik CP, Hodges JK, Nunn CL (2000) Paternity confusion and the ovarian cycles of female primates. In: van Schaik CP, Janson CH (eds) Infanticide by males and its implications. Cambridge University Press, Cambridge, pp 361–387 van Schaik CP, Pradhan GR, van Noordwijk MA (2004) Mating conflict in primates: infanticide, sexual harassment and female sexuality. In: Kappeler P, van Schaik CP (eds) Sexual selection in primates: new and comparative perspectives. Cambridge University Press, Cambridge, pp 131–150 von Engelhardt N, Kappeler PM, Heistermann M (2000) Androgen levels and female social dominance in Lemur catta. Proc R Soc Lond B 267:1533-1539

Wallen K (1990) Desire and ability: hormones and the regulation of female sexual behavior. Neurosci Biobehav Rev 14:233–241

Wallen K, Zehr JL (2004) Hormones and history: the evolution and development of primate female sexuality. J Sex Res 41:101-112

Wasser SK, Monfort SL, Southers J, Wildt DE (1994) Excretion rates and metabolites of estradiol and progesterone in baboon (Papio cynocephalus cynocephalus) feces. J Reprod Fert 101:213-220

Wasser SK, Starling AK (1988) Proximate and ultimate causes of reproductive suppression among female yellow baboons at Mikumi National Park, Tanzania. Am J Primatol 16:97-121

176

Watts DP (1990) Mountain gorilla life histories, reproductive competition, and sociosexual behavior and some implications for captive husbandry. Zoo Biol 9:185-200

Weingrill T, Lycett JE, Henzi SP (2000) Consortship and mating success in chacma baboons (Papio cynocephalus ursinus). Ethology 106:1033-1044

Wentz EA, Campbell AF, Houston R (2003) A comparison of two methods to create tracks of moving objects: linear weighted distance and constrained random walk. Int J Geogr Inf Sci 129:203-224

Wolff JO, Macdonald DW (2004) Promiscuous females protect their offspring. Trends Ecol Evol 19:127-134

Wrangham RW, Riss EVZB (1990) Rates of predation on mammals by Gombe chimpanzees, 1972–1975. Primates 31:157-170

Young WC (1961) The hormones and mating behavior In: Young WC (ed) Sex and internal secretions. Williams and Wilkins, Baltimore, pp 1173-1239 Zehr JL, Maestripieri D, Wallen K (1998) Estradiol inceases female sexual initiation independent of male responsiveness in rhesus monkeys. Horm Behav 33:95-103

Zinner D, Wertheimer J, Liedigk R, Groeneveld LF, Roos C (2013) Baboon phylogeny as inferred from complete mitochondrial genomes. Am J Phys Anthrop 150: 133-140

Zumpe D, Michael RP (1985) Mate Competition Between Female Rhesus-Monkeys. Naturwissenschaften 72:382-384

177

APPENDIX A: MALE AND FEMALE SOCIALITY MEASURES

Table A.1: Information on female dominance rank derived from Elo-ratings. Mean Female Ordinal Relative Troop Year ID Elo- Rank Rank rating AC 2012 URS 1221 1 100 UBN 1124 2 91.67 UBG 1069 3 83.33 HBN 1061 4 75 UNK 1046 5 66.67 USA 1037 6 58.33 UVA 1030 7 50 UMA 1008 8 41.67 ULY 954 9 33.33 YNN 947 10 25 YEI 889 11 16.67 YLA 844 12 8.33 YLT 771 13 0 BA 2012 WZR 1377 1 100 WAT 1261 2 94.74 WTW 1257 3 89.47 WGR 1215 4 84.21 WTO 1176 5 78.95 MAY 1111 6 73.68 ACA 1110 7 68.42 MKA 1100 8 63.16 MDA 1072 9 57.89 WDF 981 10 52.63 WOK 938 11 47.37 AEA 906 12 42.11 WTR 903 13 36.84 AKT 900 14 31.58 AKA 833 15 26.32 ATE 807 16 21.05 WLD 800 17 15.79 AJA 795 18 10.53 WDA 733 19 5.26 AMA 725 20 0 Continued

178

Table A.1 Continued

Mean Female Ordinal Relative Troop Year ID Elo- Rank Rank rating BA 2014 WGR 1187 1 100 WZR 1182 2 92.86 WAT 1163 3 85.71 WTW 1045 4 78.57 MAY 1044 5 71.43 WTO 1032 6 64.29 ATE 1029 7 57.14 AEA 1014 8 50 WDA 1000 9 42.86 WDF 971 10 35.71 WLD 919 11 28.57 ACA 911 12 21.43 AKT 909 13 14.29 AKA 799 14 7.14 AJA 795 15 0 DC 2012 SLD 1324 1 100 SER 1275 2 94.44 SAL 1143 3 88.89 HRS 1132 4 83.33 HRF 1128 5 77.78 HRT 1040 6 72.22 SIP 1025 7 66.67 SNO 1003 8 61.11 HAG 999 9 55.56 SHB 994 10 50 HOZ 966 11 44.44 SPR 956 12 38.89 HAU 926 13 33.33 SHL 901 14 27.78 SHD 880 15 22.22 SAS 852 16 16.67 HMA 851 17 11.11 HON 814 18 5.56 SAG 790 19 0 Continued

179

Table A.1 Continued

Mean Female Ordinal Relative Troop Year ID Elo- Rank Rank rating DC 2014 SER 1394 1 100 HRF 1161 2 92.86 HRT 1085 3 85.71 HRS 1065 4 78.57 SHD 1032 5 71.43 SIP 991 6 64.29 HMA 990 7 57.14 HON 954 8 50 SNO 948 9 42.86 SAS 932 10 35.71 SAG 927 11 28.57 HAU 918 12 21.43 HOZ 905 13 14.29 SHL 885 14 7.14 SPR 814 15 0

180

Table A.2: Social distance values between female-female dyads in AC group in the 2012 study period.

AC TROOP: 2012 URS UBN UBG HBN UNK USA UVA UMA ULY YNN YEI YLA YLT URS 0 UBN 97 0 UBG 152 55 0 HBN 160 63 8 0 UNK 175 78 23 15 0 USA 184 87 32 24 9 0 UVA 191 94 39 31 16 7 0 UMA 213 116 61 53 38 29 22 0 ULY 267 170 115 107 92 83 76 54 0 YNN 274 177 122 114 99 90 83 61 7 0 YEI 332 235 180 172 157 148 141 119 65 58 0 YLA 377 280 225 217 202 193 186 164 110 103 45 0 YLT 450 353 298 290 275 266 259 237 183 176 118 73 0

Table A.3: Social distance values between female-female dyads in the BA group in the 2012 study period.

BA TROOP: 2012 WZR WAT WTW WGR WTO MAY ACA MKA MDA WDF WOK AEA WTR AKT AKA ATE WLD AJA WDA AMA WZR 0 WAT 116 0 WTW 120 4 0 WGR 162 46 42 0 WTO 201 85 81 39 0 MAY 266 150 146 104 65 0 ACA 267 151 147 105 66 1 0 MKA 277 161 157 115 76 11 10 0 MDA 305 189 185 143 104 39 38 28 0 WDF 396 280 276 234 195 130 129 119 91 0 WOK 439 323 319 277 238 173 172 162 134 43 0 AEA 471 355 351 309 270 205 204 194 166 75 32 0 WTR 474 358 354 312 273 208 207 197 169 78 35 3 0 AKT 477 361 357 315 276 211 210 200 172 81 38 6 3 0 AKA 544 428 424 382 343 278 277 267 239 148 105 73 70 67 0 ATE 570 454 450 408 369 304 303 293 265 174 131 99 96 93 26 0 WLD 577 461 457 415 376 311 310 300 272 181 138 106 103 100 33 7 0 AJA 582 466 462 420 381 316 315 305 277 186 143 111 108 105 38 12 5 0 WDA 644 528 524 482 443 378 377 367 339 248 205 173 170 167 100 74 67 62 0 AMA 652 536 532 490 451 386 385 375 347 256 213 181 178 175 108 82 75 70 8 0

181

Table A.4: Social distance values between female-female dyads in the BA group in the 2014 study period.

BA TROOP: 2014 WGR WZR WAT WTW MAY WTO ATE AEA WDA WDF WLD ACA AKT AKA AJA WGR 0 WZR 5 0 WAT 24 19 0 WTW 142 137 118 0 MAY 143 138 119 1 0 WTO 155 150 131 13 12 0 ATE 158 153 134 16 15 3 0 AEA 173 168 149 31 30 18 15 0 WDA 187 182 163 45 44 32 29 14 0 WDF 216 211 192 74 73 61 58 43 29 0 WLD 268 263 244 126 125 113 110 95 81 52 0 ACA 276 271 252 134 133 121 118 103 89 60 8 0 AKT 278 273 254 136 135 123 120 105 91 62 10 2 0 AKA 388 383 364 246 245 233 230 215 201 172 120 112 110 0 AJA 392 387 368 250 249 237 234 219 205 176 124 116 114 4 0

Table A.5: Social distance values between female-female dyads in the DC group in the 2012 study period.

DC TROOP: 2012 SLD SER SAL HRS HRF HRT SIP SNO HAG SHB HOZ SPR HAU SHL SHD SAS HMA HON SAG SLD 0 SER 49 0 SAL 181 132 0 HRS 192 143 11 0 HRF 196 147 15 4 0 HRT 284 235 103 92 88 0 SIP 299 250 118 107 103 15 0 SNO 321 272 140 129 125 37 22 0 HAG 325 276 144 133 129 41 26 4 0 SHB 330 281 149 138 134 46 31 9 5 0 HOZ 358 309 177 166 162 74 59 37 33 28 0 SPR 368 319 187 176 172 84 69 47 43 38 10 0 HAU 398 349 217 206 202 114 99 77 73 68 40 30 0 SHL 423 374 242 231 227 139 124 102 98 93 65 55 25 0 SHD 444 395 263 252 248 160 145 123 119 114 86 76 46 21 0 SAS 472 423 291 280 276 188 173 151 147 142 114 104 74 49 28 0 HMA 473 424 292 281 277 189 174 152 148 143 115 105 75 50 29 1 0 HON 510 461 329 318 314 226 211 189 185 180 152 142 112 87 66 38 37 0 SAG 534 485 353 342 338 250 235 213 209 204 176 166 136 111 90 62 61 24 0

182

Table A.6: Social distance values between female-female dyads in the DC group in the 2014 study period.

DC TROOP: 2014 SER HRF HRT HRS SHD SIP HMA HON SNO SAS SAG HAU HOZ SHL SPR SER 0 HRF 233 0 HRT 309 76 0 HRS 329 96 20 0 SHD 362 129 53 33 0 SIP 403 170 94 74 41 0 HMA 404 171 95 75 42 1 0 HON 440 207 131 111 78 37 36 0 SNO 446 213 137 117 84 43 42 6 0 SAS 462 229 153 133 100 59 58 22 16 0 SAG 467 234 158 138 105 64 63 27 21 5 0 HAU 476 243 167 147 114 73 72 36 30 14 9 0 HOZ 489 256 180 160 127 86 85 49 43 27 22 13 0 SHL 509 276 200 180 147 106 105 69 63 47 42 33 20 0 SPR 580 347 271 251 218 177 176 140 134 118 113 104 91 71 0

183

Table A.7: Information on female dominance rank derived from Elo-ratings.

Troop Dates Male Mean Ordinal Relative Rank ID Elo-rating Rank Rank Trajectory AC 2012 GBS 1079 1 100.00 -0.26 LRA 1073 2 80.00 0.20 AVZ 1004 3 60.00 0.17 AMX 989 4 40.00 0.27 SHT 937 5 20.00 -0.67 SYR 918 6 0.00 0.04 0BA 2012 AST 1141 1 100.00 0.26 AGT-2 1091 2 75.00 0.89 SFI 1038 3 50.00 2.22 RED 952 4 25.00 -1.00 SIM 799 5 0.00 -0.55 BA 2014 SFI 1134 1 100.00 10.32 AST 1038 2 85.71 5.39 SUZ 1019 3 71.43 2.48 AGT-2 1003 4 57.14 -19.00 SIM 972 5 42.86 1.95 LAR 963 6 28.57 -2.66 SUR 940 7 14.29 -4.83 CRK N/A* N/A N/A N/A DC 2012 RUK 1052 1 100.00 -0.04 AMA 1047 2 87.50 0.78 HYT 1025 3 75.00 0.57 ACH 1014 4 62.50 0.53 UBO 1007 5 50.00 -2.00 LRA 1002 6 37.50 N/A BUS 963 6 25.00 -1.05 AIS 950 7 12.50 -0.30 ALA 940 8 0.00 -0.41 DC 2014 AIS 1090 1 100.00 2.45 CAN 1056 2 90.00 -0.58 HYT 1048 3 80.00 0.64 ALA 1010 4 70.00 3.19 RUK 1002 5 60.00 -3.05 A2 985 6 50.00 2.08 ARI 977 7 40.00 N/A BEAR 973 8 30.00 -1.41 LRA 960 9 20.00 N/A SUP 957 10 10.00 -0.14 MAP 942 11 0.00 -10.00

*N/A indicates information not available for that individual.

184

APPENDIX B: MALE-FEMALE ASSOCIATIONS AND FEMALE PARTNER PREFERENCES

Table B.1: Male-Female Association Indices (AIs) for group AC.

MALE AMX AVZ GBS LRA SHT SYR

UBG 0.00 0.22 0.00 0.00 0.00 0.34

ULY 0.00 -0.04 0.29 0.00 0.00 0.00 UNK 0.05 0.10 0.05 0.07 0.01 0.20

FEMALE UVA 0.01 0.00 0.13 0.40 0.02 0.01 YLT 0.12 0.00 0.02 0.00 0.13 0.45

Table B.2: Male-Female Association Indices (AIs) for group BA during the 2012 study period.

MALE

AGT-2 AST RED SFI SIM

WDF 0.00 0.35 0.00 0.00 0.29 WLD 0.04 0.00 0.00 0.00 0.71

FEMALE WTW 0.19 0.00 0.29 0.17 0.00

Table B.3: Male-Female Association Indices (AIs) for group BA during the 2014 study period.

MALE

AGT-2 AST L2 SFI SIM SUR SUZ

AJA 0.00 0.00 0.00 0.18 0.00 0.00 0.00 AKA 0.04 0.13 -0.08 0.00 0.00 0.00 0.00

FEMALE WGR 0.00 0.04 0.01 0.00 0.36 0.17 0.00

185

Table B.4: Male-Female Association Indices (AIs) for group DC during the 2012 study period.

MALE ACH AIS ALA AMA ARI BU HYT LRA MAP RUK UBO

HON 0.29 0.00 0.00 0.00 0.00 0.00 0.25 0.00 0.00 0.00 0.00

HRF 0.05 0.00 0.07 0.01 0.00 0.00 0.06 0.00 0.00 0.25 0.20 HRS 0.24 0.16 0.05 0.00 0.00 0.01 0.04 0.00 0.00 0.02 0.03

FEMALE HRT 0.06 0.00 0.00 0.44 0.00 0.00 0.00 0.00 0.00 0.00 0.00 SAS 0.01 0.02 0.08 0.00 0.00 0.17 0.02 0.01 0.00 0.38 0.00

Table B.5: Male-Female Association Indices (AIs) for group DC during the 2014 study period.

MALE A2 AIS ALA ARI BEAR CAN HYT LRA MAP RUK SUP

HAU 0.06 0.00 0.22 0.00 0.00 0.00 0.22 0.00 0.00 0.00 0.11

HRS 0.27 0.24 0.00 0.00 0.00 0.00 0.00 0.00 0.13 0.02 0.00 SAS 0.00 0.22 0.00 0.00 0.03 0.00 0.03 0.14 0.00 0.04 0.12

SER 0.00 0.16 0.24 0.00 0.00 0.26 0.00 0.00 0.00 0.00 0.00 FEMALE SPR 0.14 0.00 0.15 0.00 0.06 0.15 0.00 0.00 0.00 0.06 -0.01

186

Table B.6: Female partner preferences for Group (a) AC, (b) BA, (c) DC. (a) AC TROOP: FERTILE FEMALE 2012 UBG ULY UNK UVA YLT AMX x x x x - AVZ O x O x x GBS x O x x x

MALE LRA x x - O x SHT x x x x - SYR O x O x O (b) FEMALE FEMALE 2012 2014 WDF WLD WTW AJA AKA WGR AGT-2 X X O AGT-2 X O X AST O X X AST X O X CRK CRK X X X L2 L2 X X X

RED - - O RED MALE MALE MALE SFI X X O SFI O X X SIM O O X SIM X X O SUR SUR X X O SUZ SUZ X X X (c) FEMALE FEMALE 2012 2014 HON HRF HRS HRT SAS HAU HRS SAS SER SPR A2 A2 - O X X O ACH O - O O X ACH AIS X X O X X AIS X O O O X ALA X - - X O ALA O X X O O AMA X X X O X AMA ARI X X X X X ARI X X X X X BEAR BEAR X X X X O

BU X X X X O BU MALE MALE CAN CAN X X X O O HYT O - X X X HYT O X X X X LRA X X X X X LRA X X O X X MAP X X O X X MAP X O X X X RUK X O X X O RUK X X - X - UBO X O X X X UBO SUP SUP O X O X X

Eschewed males are marked with an (x), preferred partners are noted with a (O) and neutral males are marked with a (-).

187

APPENDIX C: FEMALE AFFILIATIONS

Table C.1: Female-female dyads, kinship, and Affiliation Indices (derived from proximity and grooming interactions) during and outside of consorts. Fem-Fem Aff. Index: Aff. Index: Dyad Kinship Group Year Consort Non-consort UBG-UN Kin AC 2012 0.00 0.18 UBG-UNK Kin AC 2012 0.60 0.06 UBG-URS Kin AC 2012 0.19 0.16 UBG-USA Kin AC 2012 0.00 0.27 UBG-UVA Kin AC 2012 0.03 0.00 ULY-UMA Kin AC 2012 0.04 0.09 ULY-UVA Kin AC 2012 0.53 0.49 UNK-USA Kin AC 2012 0.07 0.07 UVA-UMA Kin AC 2012 0.60 0.00 UVA-UN Kin AC 2012 0.17 0.12 UVA-URS Kin AC 2012 0.02 0.00 YLT-YOLA Kin AC 2012 0.00 0.00 UBG-HBN Nonkin AC 2012 0.00 0.08 UBG-ULY Nonkin AC 2012 0.03 0.00 UBG-UMA Nonkin AC 2012 0.08 0.05 UBG-YEI Nonkin AC 2012 0.03 0.06 UBG-YLT Nonkin AC 2012 0.00 0.00 UBG-YNN Nonkin AC 2012 0.00 0.06 UBG-YOLA Nonkin AC 2012 0.05 0.08 ULY-HBN Nonkin AC 2012 0.08 0.04 ULY-UN Nonkin AC 2012 0.00 0.03 ULY-UNK Nonkin AC 2012 0.07 0.04 ULY-URS Nonkin AC 2012 0.00 0.06 ULY-USA Nonkin AC 2012 0.20 0.00 ULY-YLT Nonkin AC 2012 0.04 0.13 ULY-YNN Nonkin AC 2012 0.03 0.06 ULY-YOLA Nonkin AC 2012 0.00 0.04 Continued

188

Table C.1 Continued Fem-Fem Aff. Index: Aff. Index: Dyad Kinship Troop Year Consort Nonconsort UNK-HBN Nonkin AC 2012 0.13 0.05 UNK-UMA Nonkin AC 2012 0.00 0.04 UNK-UN Nonkin AC 2012 0.00 0.22 UNK-URS Nonkin AC 2012 0.03 0.06 UNK-YEI Nonkin AC 2012 0.00 0.05 UNK-YLT Nonkin AC 2012 0.03 0.24 UNK-YNN Nonkin AC 2012 0.00 0.02 UNK-YOLA Nonkin AC 2012 0.20 0.04 UVA-HBN Nonkin AC 2012 0.02 0.08 UVA-USA Nonkin AC 2012 0.02 0.00 UVA-YEI Nonkin AC 2012 0.00 0.00 UVA-YNN Nonkin AC 2012 0.02 0.40 UVA-YOLA Nonkin AC 2012 0.02 0.04 YLT-HBN Nonkin AC 2012 0.08 0.09 YLT-UMA Nonkin AC 2012 0.00 0.02 YLT-UN Nonkin AC 2012 0.08 0.08 YLT-URS Nonkin AC 2012 0.00 0.02 YLT-USA Nonkin AC 2012 0.50 0.03 YLT-YEI Nonkin AC 2012 0.00 0.31 YLT-YNN Nonkin AC 2012 0.00 0.04 WDF-WTR Kin BA 2012 0.02 0.06 WLD-WDA Kin BA 2012 0.00 0.04 WLD-WOK Kin BA 2012 0.13 0.00 WTW-WTR Kin BA 2012 0.03 0.00 WDF-ACA Nonkin BA 2012 0.03 0.00 WDF-AEA Nonkin BA 2012 0.09 0.04 WDF-AJA Nonkin BA 2012 0.01 0.03 WDF-AKA Nonkin BA 2012 0.01 0.00 WDF-AKT Nonkin BA 2012 0.03 0.00 WDF-AMA Nonkin BA 2012 0.00 0.04 WDF-ATE Nonkin BA 2012 0.01 0.05 WDF-MAY Nonkin BA 2012 0.00 0.01 WDF-MDA Nonkin BA 2012 0.01 0.04 WDF-MKA Nonkin BA 2012 0.00 0.01 Continued

189

Table C.1 Continued Fem-Fem Aff. Index: Aff. Index: Dyad Kinship Troop Year Consort Nonconsort WDF-WATA Nonkin BA 2012 0.03 0.05 WDF-WDA Nonkin BA 2012 0.03 0.04 WDF-WGR Nonkin BA 2012 0.04 0.03 WDF-WLD Nonkin BA 2012 0.01 0.02 WDF-WOK Nonkin BA 2012 0.01 0.03 WDF-WTO Nonkin BA 2012 0.14 0.01 WDF-WTW Nonkin BA 2012 0.01 0.03 WDF-WZR Nonkin BA 2012 0.03 0.01 WLD-ACA Nonkin BA 2012 0.00 0.01 WLD-AEA Nonkin BA 2012 0.00 0.00 WLD-AJA Nonkin BA 2012 0.00 0.06 WLD-AKA Nonkin BA 2012 0.00 0.02 WLD-AKT Nonkin BA 2012 0.00 0.00 WLD-AMA Nonkin BA 2012 0.00 0.16 WLD-ATE Nonkin BA 2012 0.00 0.06 WLD-MAY Nonkin BA 2012 0.00 0.10 WLD-MDA Nonkin BA 2012 0.00 0.06 WLD-MKA Nonkin BA 2012 0.00 0.27 WLD-WATA Nonkin BA 2012 0.00 0.00 WLD-WGR Nonkin BA 2012 0.00 0.00 WLD-WTO Nonkin BA 2012 0.00 0.03 WLD-WTR Nonkin BA 2012 0.00 0.29 WLD-WTW Nonkin BA 2012 0.00 0.01 WLD-WZR Nonkin BA 2012 0.00 0.03 WTW-ACA Nonkin BA 2012 0.02 0.03 WTW-AEA Nonkin BA 2012 0.58 0.01 WTW-AJA Nonkin BA 2012 0.00 0.01 WTW-AKA Nonkin BA 2012 0.00 0.00 WTW-AKT Nonkin BA 2012 0.00 0.02 WTW-AMA Nonkin BA 2012 0.11 0.00 WTW-ATE Nonkin BA 2012 0.02 0.02 WTW-MAY Nonkin BA 2012 0.02 0.06 WTW-MDA Nonkin BA 2012 0.00 0.03 WTW-MKA Nonkin BA 2012 0.03 0.00 Continued

190

Table C.1 Continued Fem-Fem Aff. Index: Aff. Index: Dyad Kinship Troop Year Consort Nonconsort WTW-WATA Nonkin BA 2012 0.00 0.04 WTW-WDA Nonkin BA 2012 0.02 0.40 WTW-WGR Nonkin BA 2012 0.00 0.04 WTW-WOK Nonkin BA 2012 0.00 0.05 WTW-WTO Nonkin BA 2012 0.11 0.25 WTW-WZR Nonkin BA 2012 0.02 0.02 AJA-ACA Kin BA 2014 0.59 0.10 AKA-AEA Kin BA 2014 0.00 0.41 AKA-AMA Kin BA 2014 0.00 0.08 WGR-WATA Kin BA 2014 0.15 0.03 WGR-WTO Kin BA 2014 0.00 0.00 WGR-WZR Kin BA 2014 0.15 0.03 AJA-AEA Nonkin BA 2014 0.01 0.01 AJA-AKA Nonkin BA 2014 0.02 0.01 AJA-AKT Nonkin BA 2014 0.02 0.02 AJA-ATE Nonkin BA 2014 0.02 0.06 AJA-MAY Nonkin BA 2014 0.01 0.06 AJA-WATA Nonkin BA 2014 0.02 0.03 AJA-WDA Nonkin BA 2014 0.03 0.03 AJA-WDF Nonkin BA 2014 0.03 0.02 AJA-WGR Nonkin BA 2014 0.00 0.03 AJA-WLD Nonkin BA 2014 0.03 0.03 AJA-WTO Nonkin BA 2014 0.10 0.05 AJA-WTW Nonkin BA 2014 0.07 0.02 AJA-WZR Nonkin BA 2014 0.07 0.04 AKA-ACA Nonkin BA 2014 0.00 0.02 AKA-AKT Nonkin BA 2014 0.12 0.03 AKA-ATE Nonkin BA 2014 0.00 0.02 AKA-MAY Nonkin BA 2014 0.00 0.01 AKA-MDA Nonkin BA 2014 0.00 0.19 AKA-MKA Nonkin BA 2014 0.00 0.09 AKA-WATA Nonkin BA 2014 0.00 0.01 AKA-WDA Nonkin BA 2014 0.04 0.00 AKA-WDF Nonkin BA 2014 0.02 0.01 Continued

191

Table C.1 Continued Fem-Fem Aff. Index: Aff. Index: Dyad Kinship Troop Year Consort Nonconsort AKA-WGR Nonkin BA 2014 0.01 0.01 AKA-WLD Nonkin BA 2014 0.02 0.06 AKA-WTO Nonkin BA 2014 0.13 0.01 AKA-WTW Nonkin BA 2014 0.13 0.00 AKA-WZR Nonkin BA 2014 0.00 0.01 WGR-ACA Nonkin BA 2014 0.00 0.02 WGR-AEA Nonkin BA 2014 0.00 0.01 WGR-AKT Nonkin BA 2014 0.00 0.07 WGR-AMA Nonkin BA 2014 0.00 0.26 WGR-ATE Nonkin BA 2014 0.10 0.04 WGR-MAY Nonkin BA 2014 0.00 0.43 WGR-WDA Nonkin BA 2014 0.10 0.00 WGR-WDF Nonkin BA 2014 0.00 0.02 WGR-WLD Nonkin BA 2014 0.00 0.08 WGR-WTW Nonkin BA 2014 0.00 0.00 HON-HEM Kin DC 2012 0.56 0.01 HRF-HEM Kin DC 2012 0.02 0.05 HRF-HRS Kin DC 2012 0.02 0.24 HRS-HEM Kin DC 2012 0.01 0.04 HRT-HAG Kin DC 2012 0.08 0.11 HRT-HAU Kin DC 2012 0.00 0.02 HRT-HOZ Kin DC 2012 0.00 0.03 SAS-SHL Kin DC 2012 0.04 0.14 HON-HAG Nonkin DC 2012 0.00 0.04 HON-HAU Nonkin DC 2012 0.00 0.01 HON-HOZ Nonkin DC 2012 0.01 0.07 HON-HRF Nonkin DC 2012 0.08 0.01 HON-HRS Nonkin DC 2012 0.03 0.02 HON-HRT Nonkin DC 2012 0.03 0.02 HON-SAG Nonkin DC 2012 0.01 0.09 HON-SALI Nonkin DC 2012 0.01 0.04 HON-SAS Nonkin DC 2012 0.01 0.00 HON-SER Nonkin DC 2012 0.09 0.13 HON-SHB Nonkin DC 2012 0.00 0.05 Continued

192

Table C.1 Continued Fem-Fem Aff. Index: Aff. Index: Dyad Kinship Troop Year Consort Nonconsort HON-SHD Nonkin DC 2012 0.04 0.08 HON-SHL Nonkin DC 2012 0.05 0.05 HON-SIP Nonkin DC 2012 0.01 0.06 HON-SLD Nonkin DC 2012 0.01 0.02 HON-SNO Nonkin DC 2012 0.00 0.01 HON-SPR Nonkin DC 2012 0.06 0.04 HRF-HAG Nonkin DC 2012 0.00 0.22 HRF-HAU Nonkin DC 2012 0.02 0.01 HRF-HOZ Nonkin DC 2012 0.02 0.02 HRF-HRT Nonkin DC 2012 0.08 0.01 HRF-SAG Nonkin DC 2012 0.00 0.05 HRF-SALI Nonkin DC 2012 0.02 0.02 HRF-SAS Nonkin DC 2012 0.03 0.03 HRF-SER Nonkin DC 2012 0.02 0.19 HRF-SHB Nonkin DC 2012 0.05 0.01 HRF-SHD Nonkin DC 2012 0.03 0.05 HRF-SHL Nonkin DC 2012 0.03 0.03 HRF-SIP Nonkin DC 2012 0.02 0.02 HRF-SLD Nonkin DC 2012 0.03 0.01 HRF-SNO Nonkin DC 2012 0.55 0.02 HRF-SPR Nonkin DC 2012 0.03 0.04 HRS-HAG Nonkin DC 2012 0.00 0.03 HRS-HAU Nonkin DC 2012 0.00 0.01 HRS-HOZ Nonkin DC 2012 0.04 0.02 HRS-HRT Nonkin DC 2012 0.01 0.14 HRS-SAG Nonkin DC 2012 0.00 0.06 HRS-SALI Nonkin DC 2012 0.04 0.05 HRS-SAS Nonkin DC 2012 0.00 0.01 HRS-SER Nonkin DC 2012 0.01 0.05 HRS-SHB Nonkin DC 2012 0.03 0.01 HRS-SHD Nonkin DC 2012 0.01 0.03 HRS-SHL Nonkin DC 2012 0.09 0.06 HRS-SIP Nonkin DC 2012 0.01 0.01 HRS-SLD Nonkin DC 2012 0.09 0.05 Continued

193

Table C.1 Continued Fem-Fem Aff. Index: Aff. Index: Dyad Kinship Troop Year Consort Nonconsort HRS-SNO Nonkin DC 2012 0.03 0.01 HRS-SPR Nonkin DC 2012 0.01 0.04 HRT-HEM Nonkin DC 2012 0.00 0.02 HRT-SAG Nonkin DC 2012 0.25 0.03 HRT-SALI Nonkin DC 2012 0.05 0.02 HRT-SAS Nonkin DC 2012 0.02 0.01 HRT-SER Nonkin DC 2012 0.05 0.03 HRT-SHB Nonkin DC 2012 0.00 0.03 HRT-SHD Nonkin DC 2012 0.00 0.03 HRT-SHL Nonkin DC 2012 0.07 0.08 HRT-SIP Nonkin DC 2012 0.02 0.20 HRT-SLD Nonkin DC 2012 0.00 0.34 HRT-SNO Nonkin DC 2012 0.39 0.06 HRT-SPR Nonkin DC 2012 0.00 0.02 SAS-HAG Nonkin DC 2012 0.00 0.01 SAS-HAU Nonkin DC 2012 0.07 0.03 SAS-HEM Nonkin DC 2012 0.00 0.04 SAS-HOZ Nonkin DC 2012 0.04 0.09 SAS-SAG Nonkin DC 2012 0.00 0.05 SAS-SALI Nonkin DC 2012 0.04 0.01 SAS-SER Nonkin DC 2012 0.11 0.51 SAS-SHB Nonkin DC 2012 0.04 0.03 SAS-SHD Nonkin DC 2012 0.54 0.03 SAS-SIP Nonkin DC 2012 0.00 0.08 SAS-SLD Nonkin DC 2012 0.00 0.04 SAS-SNO Nonkin DC 2012 0.00 0.00 SAS-SPR Nonkin DC 2012 0.00 0.05 HAU-HOZ Kin DC 2014 0.04 0.58 HAU-HRT Kin DC 2014 0.61 0.00 HRS-HEM Kin DC 2014 0.03 0.01 HRS-HRF Kin DC 2014 0.05 0.01 SAS-SHL Kin DC 2014 0.06 0.00 HAU-HEM Nonkin DC 2014 0.00 0.01 HAU-HON Nonkin DC 2014 0.00 0.01 Continued

194

Table C.1 Continued Fem-Fem Aff. Index: Aff. Index: Dyad Kinship Troop Year Consort Nonconsort HAU-HRF Nonkin DC 2014 0.00 0.01 HAU-HRS Nonkin DC 2014 0.06 0.02 HAU-SAG Nonkin DC 2014 0.04 0.01 HAU-SAS Nonkin DC 2014 0.00 0.05 HAU-SER Nonkin DC 2014 0.17 0.00 HAU-SHL Nonkin DC 2014 0.04 0.01 HAU-SIP Nonkin DC 2014 0.00 0.05 HAU-SNO Nonkin DC 2014 0.00 0.04 HAU-SPR Nonkin DC 2014 0.05 0.05 HRS-HON Nonkin DC 2014 0.04 0.05 HRS-HOZ Nonkin DC 2014 0.02 0.03 HRS-HRT Nonkin DC 2014 0.62 0.01 HRS-SAG Nonkin DC 2014 0.04 0.00 HRS-SAS Nonkin DC 2014 0.02 0.06 HRS-SER Nonkin DC 2014 0.07 0.02 HRS-SHL Nonkin DC 2014 0.01 0.01 HRS-SIP Nonkin DC 2014 0.02 0.05 HRS-SNO Nonkin DC 2014 0.01 0.02 HRS-SPR Nonkin DC 2014 0.01 0.02 SAS-HEM Nonkin DC 2014 0.04 0.02 SAS-HON Nonkin DC 2014 0.09 0.07 SAS-HOZ Nonkin DC 2014 0.06 0.02 SAS-HRF Nonkin DC 2014 0.09 0.06 SAS-HRT Nonkin DC 2014 0.56 0.02 SAS-SAG Nonkin DC 2014 0.02 0.05 SAS-SER Nonkin DC 2014 0.00 0.02 SAS-SHD Nonkin DC 2014 0.00 0.01 SAS-SIP Nonkin DC 2014 0.00 0.31 SAS-SNO Nonkin DC 2014 0.04 0.30 SAS-SPR Nonkin DC 2014 0.04 0.05 SER-HEM Nonkin DC 2014 0.01 0.01 SER-HON Nonkin DC 2014 0.01 0.00 SER-HOZ Nonkin DC 2014 0.06 0.12 SER-HRT Nonkin DC 2014 0.01 0.45 Continued

195

Table C.1 Continued Fem-Fem Aff. Index: Aff. Index: Dyad Kinship Troop Year Consort Nonconsort SER-SAG Nonkin DC 2014 0.02 0.08 SER-SHD Nonkin DC 2014 0.01 0.03 SER-SHL Nonkin DC 2014 0.12 0.02 SER-SIP Nonkin DC 2014 0.01 0.02 SER-SNO Nonkin DC 2014 0.00 0.01 SER-SPR Nonkin DC 2014 0.10 0.13 SPR-HEM Nonkin DC 2014 0.03 0.01 SPR-HON Nonkin DC 2014 0.02 0.02 SPR-HOZ Nonkin DC 2014 0.00 0.04 SPR-HRF Nonkin DC 2014 0.08 0.27 SPR-HRT Nonkin DC 2014 0.63 0.42 SPR-SAG Nonkin DC 2014 0.03 0.02 SPR-SHD Nonkin DC 2014 0.01 0.00 SPR-SHL Nonkin DC 2014 0.03 0.03 SPR-SIP Nonkin DC 2014 0.01 0.04 SPR-SNO Nonkin DC 2014 0.01 0.00

196