SOCIAL FACTORS AFFECTING FEMALE BEHAVIOR, ECOLOGY, AND FITNESS IN WILD BOTTLENOSE DOLPHINS

Transcription

1 SOCIAL FACTORS AFFECTING FEMALE BEHAVIOR, ECOLOGY, AND FITNESS IN WILD BOTTLENOSE DOLPHINS A Dissertation submitted to the Faculty of the Graduate School of Arts and Sciences of Georgetown University in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Biology By Megan Marie Wallen, B.S. Washington, DC December 2, 2016

2 Copyright 2016 by Megan Marie Wallen All Rights Reserved ii

3 SOCIAL FACTORS AFFECTING FEMALE BEHAVIOR, ECOLOGY, AND FITNESS IN WILD BOTTLENOSE DOLPHINS Megan Marie Wallen, B.S. Thesis Advisor: Janet Mann, Ph.D. ABSTRACT In humans and non-human animals alike, the social environment presents contextdependent costs and benefits. Some species face extreme pressures resulting from divergent fitness traits among the sexes, leading to pronounced sexual conflict. In this dissertation, I capitalize on the 30-year longitudinal dataset on Shark Bay bottlenose dolphins (Tursiops aduncus) to understand how the fission-fusion society facilitates preferred associations and emergence of social network properties that influence fitness. Specifically, I investigate how the coercive mating system influences intersexual associations and imposes ecological costs from a perspective not often taken that of the adult female. Social dynamics impact female reproduction, and may serve to mitigate consequences of male harassment. Chapter 1 characterizes patterns of intersexual association near conception. Females increase association with adult males just prior to conception, but juvenile male association always remains low. As expected, intersexual associations dropped off significantly postconception, demonstrating males could detect pregnancy fairly early. Surprisingly, despite a system of bisexual philopatry whereby males and females remain in their natal area for life, mothers and weaned sons did not appear to exhibit avoidance patterns, except during the mothers fertile periods. Chapter 2 quantifies the ecological costs to females when in the presence of adult males, who present a threat of sexual coercion. Females appear to suffer an ecological cost to increased iii

4 male association, in that preferred ranging and habitat use patterns are constrained. However, the interaction between male impact and female reproductive status demonstrates that cycling females experience the greatest impact. Finally, Chapter 3 explores the early differentiation of social tactics in juvenile female dolphins, and quantifies the relationship between social network measures and fitness. Several metrics predict calving success, but surprisingly network centrality is not a strong influence. These results suggest that strong, differentiated dyadic bonds, likely with kin, are beneficial during development and aid in later reproductive success. This work contributes to our understanding of how animals, females in particular, exploit and mitigate the unique social pressures of a coercive mating system, and highlights the importance for future studies to quantify the fitness benefits and consequences of social relationships in a free-ranging, socially complex mammal. vi

5 Coffee: this would not have been possible without you. Much appreciated, MEG vii

6 ACKNOWLEDGEMENTS Though words can never fully express my gratitude for the support I ve received throughout this life-changing process, I will do my best to acknowledge those who have enhanced my life, my being, and by extension this final product of my work. Janet I will forever be grateful for the world of opportunity you have provided me simply by taking on the challenge of having me as your grad student. There were some good moments, and some learning moments, but your belief in me persevered, therefore I persevered. I admire your critical eye, your ruthless devotion, your trust in your students. I hope to emulate those qualities someday. Thank you to my committee members Peter Armbruster, Shweta Bansal, and Carson Murray for providing thoughtful and constructive feedback along the way. Peter, thank you for enduring all three of my committees and defenses over the years. You have a way of verbalizing difficult concepts and that has been very valuable to my education. Shweta, thank you for always asking the important questions; your positive attitude always kept me going. Carson, your clarity on the big picture issues transformed my work. Through the tears, you all have made me a better scientist. My work would not have been possible without the dedication of many assistants throughout the years. Mother Bird and Mr. Muscles as three first-timers together in Monkey Mia, we were quite the trio. Cayenne pepper will never taste the same. Late night chats. Des and Doris, the Little Mann Van you had to deal with the trials and tribulations of my first time leading a field season independently. The bar is one that can be tripped over, that s for sure. I will never forget the Presidential Photo, the Tzatziki Pail, animal movies, and baking a dessert viii

7 for 4 hours straight. What a beautiful mess we were. Thank you for your patience. DT4L, Anni and Sara, our relationship as close friends and colleagues made the season a productive one. Thank you for the fun, laughs and your hard work. Anni & Jillian, our season started off with a few bumps in the road, but you were troopers. I am so grateful for the hard work you both put in, and Jillian, as a first-timer in remote WA you handled everything with grace. I want to give a special shoutout to Anni my assistant 4 lyf, as well as friend for life. She has been with me for each of my field seasons, countless months back in the lab at Georgetown, and now on home soil for the Potomac project. You are someone I know I can always count on, and I m excited to see where your career takes you and be a part of that process. Viv you weren t my assistant, but as an & Co member throughout my tenure, you have provided me with advice and insight for which I am truly thankful. I am so grateful for all of the hard work my assistants put in. It s not always easy to live and work in close quarters in a remote location for extended periods of time, but you all made my job easy. I have learned as much from you as I have from the luminaries. I would like to thank all of my friends and colleagues who I have met in Australia. Aspen Parks thank you for supporting me and allowing me to conduct research in your beautiful home. Thank you to all of the DPaW rangers for your enthusiastic curiosity, passion for the dolphins, and dedication to their welfare. You opened my eyes to the realities of wildlife conservation. Also, thank you to Harvey, the whole Shotover crew, and members of SBERP. On a personal level, I have to thank Grant. Roundy, I m without words. I never would have made it without you. You re a best-friend-father-figure, and I miss you every day. I so enjoyed coming back year after year; you were a constant presence that made Shark Bay feel like home. Watching you build your house over the years, fishing in the dark, fancy dinners, bloodies after work, pizza night, ALL the dinner nights, Coral Bay trip, beach bonnies, Mad Max nights, are ix

8 among my best memories in WA. You re a true Aussie. Wayno you smile like Santa and talk like a truck driver, and I hope you never change; you never failed to put a smile on my face. Claudia, beautiful Claudia, you gave me inspiration to seek drive with contentment; a difficult balance. Your energy is contagious. And finally, I can t end this section without acknowledging the Bay Bogans. You guys, what do I even say. Bangers for breakfast. Emu for lunch. You have not realized it but through the ups and downs of our brief time together you all have truly taught me about myself. It is a learning that never stops and for that I am ever grateful. The Mann Lab rs will always hold a dear space in my heart for all of the camaraderie we ve had over the years. Mags, I have always admired your intellect and steadfastness. I feel honored to be working on a paper with you. Eric Conserve and Protect Patterson, I would never have made it through without your advice, stats help, and ability to be objective and bring perspective to any situation. Ev, lady, you ve been a teacher, a mentor, but most importantly a great friend. I m secretly glad you keep coming back each year, and that you ve been by my side up through my defense! Excited to join the postdoc world with you! Caitlin, you ve been an amazing friend over the years. I love that we can have intellectual conversations over beers, or celeb gossip during work breaks. Madison, we ve become a lot closer over the past year and I m so grateful for your friendship. You re always a positive presence that helps me gain a better outlook on things. Taylor and Jared you guys brought a fresh energy and perspective that made my final months endurable. My cohort buddies Ricardo, Xin, and Shu kept me going through the hard times. Ricardo you ve been by my side, commiserating and celebrating together over the past 5+ years. I m so grateful for your friendship. And my non-georgetown friends: Shannon and James, x

9 JWhite, Jennifer and Chris (and Maizey!), Ali, Ana, and Princess. You all have been supportive bystanders as I ve gone through the ups and downs of my career. Most importantly, I am forever indebted to my eternally supportive family. Jeff, Kellie, Mom, and Dad, thank you for your unending support and enthusiasm as I ve pursued my unconventional career. It has meant so much to have you guys cheering for me along the way. I m so glad you all came out to share in my defense weekend. Also, Mom and Dad, you may not admit it but I think I have inspired you guys to start your own research project right at home. Between the deer, coyotes, black squirrels, fox squirrels, red squirrels (Mom you should know the difference by now!), turkeys, fawn birth, raccoons mating in a tree, and robin-antagonizing, I have never received so much news about animals inhabiting our own backyard. Over the years your analysis of their behavior has grown more sophisticated. Time to start point sampling! And last but certainly not least, I want to send out a thank you to all animals, everywhere, for giving me the curiosity to observe and learn about your fascinating behaviors. From squishy dolphins to greedy hyaenas, you ve illuminated my understanding of the earth and my own connection to nature. I will do my best to make sure your world stays as untouched and pristine as it should be, allowing the cycle of life, death, and evolution to occur, just as it should as nature intended. {} xi

10 TABLE OF CONTENTS INTRODUCTION... 1 CHAPTER I. INTERSEXUAL ASSOCIATIONS IN BOTTLENOSE DOLPHINS: FEMALES ASSOCIATE WITH PRIME MATES BUT NOT SONS DURING ESTROUS... 5 Introduction... 5 Methods... 9 Study site... 9 Data collection life history Data collection association measured by surveys Data analysis Results Female-male association patterns around conception Mother-son association patterns Discussion Figures and Tables CHAPTER II. THE ECOLOGICAL COSTS TO FEMALES IN A SYSTEM WITH ALLIED SEXUAL COERCION Introduction Methods Study population and site Data collection Data analysis xii

11 Results Female associations with males Ranging Habitat use Discussion Figures and Tables CHAPTER III. THE FITNESS CONSEQUENCES OF EARLY SOCIAL NETWORK STRATEGIES IN WILD BOTTLENOSE DOLPHINS Introduction Methods Study site Data collection life history and reproduction Data collection survey association Data analysis social network metrics Data analysis social strategies and fitness Results Discussion Figures and Tables CONCLUSION REFERENCES Introduction Chapter I xiii

12 Chapter II Chapter III APPENDIX A: Chapter II Supplemental Material Methods Single male associations Results Ranging Habitat use Discussion Figures and Tables APPENDIX B: Publishing Note APPENDIX C: Chapter III Supplemental Material Figures and Tables APPENDIX D: Acknowledgments Chapter I Chapter II Chapter III xiv

13 LIST OF FIGURES Figure I.1a. The fitted quadratic model with 95% confidence intervals for the predicted number of adult males associating with females by day from conception Figure I.1b. The fitted quadratic model with raw data points and 95% confidence intervals for the predicted number of adult males associating with females by day from conception Figure I.2. Box and whisker plot showing the number of adult males associating with females by day from conception Figure I.3a. The fitted linear model with 95% confidence intervals for the predicted number of juvenile males associating with females by day from conception Figure I.3b. The fitted linear model with raw data points and 95% confidence intervals for the predicted number of juvenile males associating with females by day from conception Figure I.4. Proportion of mother-son pairs seen together by son s age in years post-weaning. 29 Figure I.5. Predicted probability with 95% confidence interval of mother-son pairs being sighted together post-weaning, given son s age Figure I.6. Proportion of time that females were with adult or juvenile males when she was cycling Figure II.1. Adult female (N = 32) (a) distance from the baseline centroid, (b) average ranging distance, and (c) primary habitat selection ratios for each combination of cycling and male presence factor levels Figure II.2. Adult male (N = 73) (a) distance from the baseline centroid, (b) average ranging distance, and (c) primary habitat selection ratios for each factor level Figure III.1. Effect of normalized degree on calving success Figure III.2. Effect of normalized strength on calving success Figure III.3. Effect of clustering coefficient on calving success Figure III.4. Biplot displaying each female (N = 33) in the principal component space and the loadings of each metric on each principal component Figure III.5. Effect of PC1 on calving success xv

14 Figure A.S1. Adult female (N = 11) (a) distance from the baseline centroid, (b) average ranging distance, and (c) primary habitat selection ratios for the subset female sample analysed with a single male present for each combination of cycling and male presence factor levels Figure C.S1. Each principal component on the x-axis (1-5) with the cumulative variance explained Figure C.S2. Plot of each of the 33 juvenile females in their relative locations in the principal component space Figure C.S3. Plot of each social network metric relative to its weight on each principal component xvi

15 LIST OF TABLES Table I.1. Parameter estimates for fixed effects in the quadratic regression model for adult male association Table I.2. Parameter estimates for fixed effects in the linear regression model for juvenile male association Table II.1. Proportion of sightings per individual per factor level used in the full data set Table II.2. MANOVA results for overall female habitat use selection ratios and the effect of female cycling status and male presence, and overall male habitat use selection ratios and the effect of female presence including cycling status Table III.1. Description of each social network metric analysed in this study, with a comprehensive compilation of studies that have found the metrics to predict direct measures of fitness (including survival and reproduction) or other variables related to fitness Table III.2. Parameter estimates for fixed effects for individual linear regression models Table III.3. Coefficients of the linear combinations of each continuous variable for the first three principal components Table III.4. Relative importance of the first three principal components and proportion of variance explained by each principal component Table III.5. Parameter estimates for fixed effects for the linear regression model on the principal components Table A.S1. MANOVA results for overall female habitat use selection ratios and the effect of female cycling status and male presence Table C.S1. Matrix of correlation coefficients between each social network metric xvii

16 INTRODUCTION Group living has many well-known adaptive benefits for animals, allowing the forging of social bonds, efficient information transfer, reduced predation risk, and cooperative actions (Alexander, 1974). However, with increasing social complexity (Freeberg et al., 2012), social relationships can become a significant source of physiological and social stress, leading to potentially negative consequences. Much research has focused on the costs and benefits of social relationships, yet the mechanisms by which social relationships affect fitness appear to be context dependent. Thus, humans and non-human animals alike have evolved to exploit and mitigate the pressures of the social environment. Such responses, in turn, have important implications for behavioral patterns, fitness, and ultimately the evolution of societies. In the animal kingdom, the evolution of varied social systems from solitary to familial to group living, from monogamy to high promiscuity has led to the evolution of numerous sources of social interaction, and therefore many potential sources of social conflict. An interesting case is the fission-fusion society, in which group size and composition are ephemeral and change frequently. Such fluidity allows for the emergence of individual variation in association preferences and avoidances, and those that are highly context-dependent. This dissertation focuses on two aspects of the adaptive consequences of social relationships in a fission-fusion society. First, I focus on male-female conflict with respect to reproduction, or sexual conflict. Sexual conflict arises when fitness traits of males and females are highly divergent (Bateman, 1948; Arnqvist and Rowe, 2013). With high reproductive investment, females are choosier about their mates and resist additional costly matings (Bateson, 1983). In contrast, males enforce a higher reproductive rate with minimal energetic investment 1

17 most pronounced in mammals. Each sex therefore has evolved strategies to overcome the interests of their mating partner and to drive reproduction towards their own interests (Andersson, 1994). One male strategy to overcome female resistance to mating is sexual coercion, whereby males use force, or the threat of force, to increase the chance that a female will mate with that particular male, and decrease the chance that she will mate with other males (Smuts and Smuts, 1993; Clutton-Brock and Parker, 1995; Watson-Capps, 2009). Females may suffer direct fitness consequences as a result of male coercion (i.e., reduced survival and reproduction). However, the consequences of coercion, or the threat of coercion, may also be subtle, complex, and/or delayed, particularly in long-lived, socially dynamic, and ecologically complex species. Such consequences may also have significant impact on fitness measures. The coercive mating system also presents an interesting case in which to study social counterstrategies from the female perspective. Thus, my second focus is the adaptive consequences of female social relationships, specifically during development, prior to sexual maturity. In mammals with long lifespans, extended developmental periods, social complexity, behavioral diversity, and ecological specialization, long-term fitness consequences are more difficult to study, given the extensive data needed over an individual s lifetime. To answer these questions, I capitalize on the 30-year longitudinal dataset on Shark Bay bottlenose dolphins (Tursiops aduncus), a long-lived mammal with a coercive mating system. Adult males form stable alliances and cooperate to consort with females to gain mating access, sometimes involving extreme aggression and harassment (Connor and Krützen, 2015). There are already several documented costs of male coercion to females, such as physical injury (Scott et al., 2005) and a reduced foraging budget (Watson-Capps, 2005). Further, Shark Bay dolphins 2

18 present a rare case of bisexual philopatry (Tsai and Mann, 2013), whereby both sexes remain in their natal areas throughout their lifetime. Many animals avoid the risk of inbreeding depression by single-sex dispersal (Wrangham, 1980; Clutton-Brock and Lukas, 2012) or kin recognition (Mateo, 2004). However, having lifelong overlapping home ranges enables animals to form and develop social relationships from an early age, often with inclusive fitness benefits. In Chapter 1, I examine patterns of intersexual associations as a function of female reproductive state. Given preferred same-sex associations in this population (Mann et al., 2012), I expect adult female-adult male associations to be primarily restricted to estrus periods near conception. Importantly, because neither male nor female Shark Bay dolphins disperse, I examine mother-son associations post-weaning, and test the hypothesis that there would be clear avoidance patterns, given that inbred females have reduced fitness (Frère et al., 2010). Shark Bay dolphins exhibit individual variation and specialization in foraging tactics (Mann and Sargeant, 2003; Mann et al., 2008; Sargeant and Mann, 2009) and related habitat use (Sargeant et al., 2007; Patterson and Mann, 2011). Individual preferences are socially learned at a young age from the mother, and are typically female biased. Given that foraging ecology requires a fair amount of skill development even into adulthood (Patterson et al., 2016), a disruption of foraging budgets during consortships might have fitness consequences. In Chapter 2 (Wallen et al., 2016), I test the hypothesis that adult males, who present a coercive threat when females are cycling, influence female ranging and habitat use important predictors of foraging ecology. I predict fitness variation among Shark Bay females may be due in part to the ecological impacts of the threat of coercion. Finally, social patterns in female Shark Bay dolphins are also highly individualized. Among males, alliance formation and membership is crucial for successful reproduction 3

19 (Krützen et al., 2004). However, female social bonds are much more variable, with some individuals being highly gregarious, while others are relatively solitary (Smolker et al., 1992; Mann et al., 2000; Gibson and Mann, 2008a; b; Mann et al., 2012). It is still a conundrum why females do not form strong bonds as a rule, as in other socially complex, long-lived species with female philopatry such as baboons (Papio cynocephalus; Silk et al., 2009; 2010). In Chapter 3, I explore social strategy differentiation and reproductive fitness consequences by taking a social network approach. Previous studies have shown the importance of taking this approach to understand calf development (Stanton et al., 2011), and early-age survival (Stanton and Mann, 2012). Females may benefit from various types of dyadic social bonds (Seyfarth and Cheney, 2012), or by network connectedness (Cheney et al., 2016). I hypothesize that social bonds may be one mechanism by which females mitigate the pressures of sexual coercion and other types of social competition. By fostering social support from kin, close associates, and/or the wider social network, particularly from a young age, females may demonstrate individuality in the specific benefits gained from social relationships. Intrapopulation fitness variation in mammalian females is important for species ecology, evolution, and population dynamics. I hypothesize that the various ways the social environment influences individual female behavior and ecology also impacts fitness. With sophisticated methods focused on fitness consequences of behavior and sociality in wild populations, the field of behavioral ecology has the potential to inform questions related to population viability and ultimately conservation. Most notably, this work highlights the importance of intraspecific variability in responses to social stress and strategies to navigate social pressures that impact female reproduction. 4

20 CHAPTER I Intersexual associations in bottlenose dolphins: females associate with prime mates but not sons during estrous 1 INTRODUCTION Intersexual association patterns are expected to vary with female reproductive status. Given their greater investment per offspring, mammalian females need to discriminate and avoid suboptimal mates, including subadult males and relatives. In many species, subadult males are subprime mates for a variety of reasons, including lower physiological reproductive potential (e.g. Yuen et al., 2007), subordinate status and lower competitive ability (Dobson, 1982). Close kin are also suboptimal mating partners. In most plant and animal species, offspring produced by mating between close relatives suffer reduced fitness, or inbreeding depression (Darwin 1868; 1876; Álvarez et al., 2015). Most mammalian societies have mechanisms that reduce the probability of mating with a close relative, including male-biased dispersal (Clutton-Brock and Lukas, 2012), or more rarely, female-biased dispersal (e.g. chimpanzees, Pan troglodytes; Pusey and Packer, 1987), such that potential mating partners are not likely to be direct relatives (Pusey and Wolf, 1996). Inbreeding can also be avoided in stable societies through kin recognition (Mateo, 2004), via prior association familiarity, or mechanisms of phenotype matching (e.g., detection of MHC dissimilarity: Brown and Eklund, 1994; Penn, 2002). 1 Authorship for paper: Megan M. Wallen, Ewa Krzyszczyk, & Janet Mann. 5

21 In fission-fusion social organizations, grouping patterns are temporally and compositionally variable as subgroups change size and composition dynamically (Aureli et al., 2008). The social fluidity of fission-fusion species therefore provides an insightful paradigm in which to consider how females navigate the costs and benefits of association patterns during conceptive periods. This is particularly true in species such as bottlenose dolphins (Tursiops spp.), which are characterized by a high degree of fission-fusion dynamics. In this study we investigate context-dependent association patterns that emerge as a result of social fluidity. Bisexually philopatric species present an unusual case in the animal kingdom in which both sexes remain in their natal area after reaching sexual maturity. Here, animals have the added complexity of searching for mating opportunities while reducing the consequences of close inbreeding (Frère et al., 2010a). In such species, mechanisms of kin recognition or avoidance are imperative for appropriate mate selection. For example, both sexes of fish-eating killer whales (Orcinus orca; Baird and Whitehead, 2000; Wright et al., 2016), and long-finned pilot whales (Globicephala melas; Amos et al., 1993; Ottensmeyer and Whitehead, 2003) stay in their natal pod for life, but mating occurs outside of their primary social group. Here, kin recognition is obvious because the group membership is highly stable, and in killer whales, closely related matrilineal units maintain similar dialects (Deecke et al., 2010). In bottlenose dolphins where fission-fusion dynamics are high and group composition changes on a minute-by-minute basis, bisexual philopatry takes a different form, where sons and daughters maintain locational philopatry but daughters maintain higher matrilineal social philopatry (Tsai and Mann, 2013). Thus, we expect intersexual associations to reflect a balance between female association with prime males and reduced association with closely related males during conceptive periods. 6

22 An added driver of intersexual associations is the threat of infanticide. In promiscuous species, multiple mating occurs even after pregnancy, presumably to confuse paternity and thereby reduce the risk of infanticide (van Schaik et al., 2004). Females may tolerate prolonged male presence or additional matings, and as such intersexual associations may not reliably match up with female reproductive state. In this study, we investigate how female bottlenose dolphins navigate the conflicting demands of mate selection while reducing the risk of inbreeding with close relatives in Shark Bay, Western Australia. This population is characterized by a fission-fusion society, whereby foraging is primarily a solitary activity (Mann and Sargeant, 2003; Sargeant et al., 2007; Sargeant and Mann, 2009), but groups form during socializing, travelling, some types of foraging, and resting (Heithaus and Dill, 2002; Gero et al., 2005). During the breeding season and presumably during female cycling periods, males of one or more alliances cooperate to consort with females for mating access and often coerce consortships with conspicuous aggression (Connor et al., 1992a; b; Connor et al., 1996). Indeed, Scott et al. (2005) found that cycling females had greater tooth rake injury compared to non-cycling females. Tooth rakes are signs of conspecific aggression, and their presence indicates greater male association (e.g. consortships) during that time because females are virtually never aggressive (Scott et al., 2005). For females with nursing offspring, consortships tend to occur when her calf is within a year of being weaned, and females that were consorted in one year were more likely to give birth the following year than females who were not consorted (Connor et al., 1996), suggesting that association patterns with adult males may be a reliable indicator of female cycling. Previous work in Shark Bay has shown that male-female association patterns reflect female reproductive state, with intersexual dyadic association being highest when females are 7

23 cycling compared to when they are pregnant (Smolker et al., 1992). Importantly, despite the opportunity for frequent contact with conspecifics, male-female association is virtually always low: adult females spend less than 20% of their time with adult males outside of cycling periods (Wallen et al., 2016), and less than 5% of all male-female pairs seen together have preferred associations, compared to 59% of all female-female pairs, and 85% of all male-male pairs (Smolker et al., 1992). Also, mothers with dependent calves associate with juvenile and adult males less than expected based on their availability (Gibson and Mann, 2008b). Further, Shark Bay dolphins preferentially associate with kin compared to non-kin (Frere et al., 2010b). However, upon weaning, mother-offspring association, and particularly motherson association, declines dramatically despite the lack of geographic or social dispersal in this population (Tsai and Mann, 2013). Social avoidance post-weaning may be one mechanism by which mothers and sons reduce the risk of inbreeding. Reduced mother-son association may also indicate the time when males reach physiological or behavioral (e.g. alliance formation and competitive ability) sexual maturity. Alternatively, because females with dependent offspring generally do not associate with juvenile males (Gibson and Mann, 2008b), it could be that the fact that her son is a juvenile rather than kinship driving the reduction in association. Here we seek to understand female association patterns with respect to reproductive status, male age and relatedness. Specifically, we hypothesize that females optimize mate choice directly or indirectly by associating preferentially with adult males, and reducing association with juvenile males and sons near conception. Second, if infanticide presents a real risk to females as proposed previously (Connor et al. 1996), then females are expected to continue association with adult males post-conception. We predict that adult female association is i) highest with adult (prime) males around conception, and ii) lowest with juvenile (subprime) 8

24 males around conception given their subordinate status. If adult male-female association is strongly dictated by conception date, this would suggest that males can detect fertile periods with a high degree of accuracy. This would undermine the infanticide argument. We also expect that kin (inbreeding) avoidance will be apparent for mother-son pairs, despite bisexual philopatry (Tsai and Mann, 2013) and evidence for inbreeding in the population (Frère et al., 2010a). We predict that mothers will associate less with their weaned sons iii) as they transition from the juvenile to the adult stage, iv) in comparison to unrelated juvenile or adult males, and v) during cycling (fertile) periods. Finally, when mothers and adult sons are together, we hypothesize that inbreeding risk may be mediated by group dynamics. Specifically, we predict that vi) group size will be larger with adult sons compared to juvenile sons, providing a social buffer, and vii) group activity state will primarily be resting or traveling (non-social contexts) when females are with their adult sons compared to with their juvenile sons. METHODS Study site Shark Bay, Western Australia (25 47 S, E) is home to a resident population of bottlenose dolphins (Tursiops aduncus) that has been studied continuously since Researchers have collected behavioral, ecological, genetic, and demographic data on > 1600 individuals within the 300 km 2 study area on the eastern gulf near Monkey Mia. In particular, mothers and calves have been intensively tracked, enabling detailed study of female association 9

25 patterns over time. Matrilineal relationships are known through observed association with the mother early in life. Individuals are uniquely identified by the shape and damage to the dorsal fin using photo ID, as well as other obvious bodily markings such as tooth rakes and shark bites (Whitehead et al., 2000). Sexes are determined by views of the genital area, consistent presence of a dependent calf (Smolker et al., 1992; Mann et al., 2000), and genetics (Krützen et al., 2004). Data collection life history For this study, adult females were included if they had at least one calf with a highly accurate birthdate (± 3 days) and were observed at least once within eight months of conception (N = 65 adult females, 104 calves). Only calves whose birthdate assignment was accurate to the day (±3 days) were used to determine conception. Calves can be aged precisely by both physiological and behavioral characteristics (fetal folds, floppy dorsal fin, curled tail flukes, cork-like surfacings), in addition to sightings of the mother just before and just after the birth (Mann and Smuts, 1999). Conception date was assigned based on the 12-month gestation period documented for captive Tursiops spp. (Schroeder, 1990; Lacave et al., 2004). To investigate how associations with sons and unrelated males change during the receptive period, we classified males as juveniles (subprime) or adults (prime). The juvenile period is defined as the period between the weaning age (average is 4 yrs; Mann et al., 2000) until the age of 10. In the Shark Bay population, the earliest pregnancy occurs at age 10 (J.M., unpublished data), and male bottlenose dolphins in captivity are almost always sexually mature by age 10 based on testosterone levels (Wells et al., 1987; Brook et al., 2000; Yuen, 2007) even 10

26 if asymptotic growth has not been reached (Read et al., 1993). Thus, males were considered to be in the adult age class from the age of 10 onwards. Data collection association measured by surveys Female association data were calculated from 1211 unique surveys conducted from 1989 to 2015 (N = 65 adult females, 104 conceptions). Only one survey per female per day was selected to reduce temporal autocorrelation. Surveys are brief, five-minute snapshots of group behavior and composition collected by scan sampling (Altmann, 1974; Mann, 1999). Predominant group behavior was defined as the activity of 50% of group members during the first 5-minutes of the survey. Activities included foraging, resting, socializing, and traveling (see Karniski et al., 2015 for an ethogram). Dolphins were considered to be associating (group members) if they were within 10 meters of any dolphin in the group during the survey (Smolker et al., 1992). To investigate how female association changes around receptivity, we considered all survey data in which a female was observed four months before conception (when the female is cycling) versus four months after conception (when the female is pregnant). Four months was selected to encapsulate periods prior to females becoming attractive, and after conception when female attractivity is expected to wane. All metrics were calculated separately for each female to account for individual differences in gregariousness; females may have different baseline levels of association with males regardless of reproductive state. We first calculated the average number of males each female was found associating with across all surveys in which she was observed. such that each 11

27 individual female was assigned her own male average across all surveys. The number of males in each survey during the pre- and post-conceptive time periods was then subtracted from each female s male average to obtain her male difference, which is the number of males present in a female s survey relative to that female s male average. Therefore, any values greater than 0 indicate higher than average association, while any values less than 0 indicate lower than average association. We calculated a male difference for adult male association and juvenile male association separately. To test how mother-son association varies by reproductive status, we included mothers that were known to be alive while their sons were juveniles (N = 59 mothers, 78 sons) and/or adults (N = 40 mothers, 44 sons). Each mother-son pair had to have 15 surveys or more during the offspring s respective juvenile or adult time period. We compared mother-son association to mother-unrelated male association, both at the juvenile and adult stages. Unrelated male association was calculated as the number of surveys (excluding those with sons) with juvenile or adult males, divided by the number of available juvenile or adult males in our entire study population during that time period, to get the average number of surveys per unrelated male. Because the age at which males become fertile isn t well defined (between 8-13 years depending on species (T. aduncus or T. truncatus) and whether captive or not see Read et al., 1993), we further classified sons by year of age, and calculated the proportion of mother-son pairs that were seen associating at least once within each son age-year class. Finally, we compared group size and group activity when mothers were with their juvenile sons versus with their adult sons. 12

28 Data analysis We used permutation tests to compare the proportion of surveys in which a female associated with juvenile (subprime) and adult (prime) males in the two conception windows. To assess juvenile and adult male association from pre- to post-conception on a continuous scale, we used linear mixed effect models with day from conception as a fixed effect and the calf s ID as a random effect (already nested within the mother s ID). Models for juvenile and adult males were run separately as association was on vastly different scales (mean juvenile male average: 0.34; mean adult male average: 1.34). Exploratory analysis indicated that a quadratic function was a better fit for adult male association, while linear was the best fit for juvenile male association. We calculated of R-squared for goodness of fit for mixed effect models following Xu (2003). To determine the breakpoint at which female association patterns with adult males changed, we ran a piecewise linear regression using fixed effects only (R package segmented; Muggeo, 2003; 2008). We first specified the single continuous variable on which to determine a segmented relationship (day from conception), and provided a start value for the breakpoint estimation based on a visual examination. This method uses maximum likelihood to calculate the distance between each segment and converges when the distance is 0, making the model nearly continuous. We tested the null hypothesis that the difference in slope on either side of the breakpoint is zero. To investigate how kinship may influence association, we first used permutation tests to compare son versus unrelated male association, separately at the juvenile and adult stages. We then refined age class by year and ran a logistic regression to calculate the probability of a weaned son being seen with his mother given his age, using age in years as a fixed effect and the 13

29 mother s ID as a random effect. To compare group size when mothers were with their juvenile versus adult sons, we ran a generalized linear mixed-effect model with a Poisson distribution. The son s age class (juvenile or adult) was included as a fixed effect and the mother s ID was included as a random effect. All statistical analyses were run in R v3.3.1 (R Core Team, 2016). RESULTS Female-male association patterns around conception Adult females associate with more adult males than juvenile males both before and after conception (pre-conception: t = -7.47, df = 1816, P < ; post-conception: t = -4.33, df = 757, P < ). Female association with adult males showed a negative quadratic relationship across the conception window, with highest association just prior to and at conception (Table I.1; Figures I.1a,b; 2). The polynomial quadratic model fit significantly better than the linear model excluding the quadratic term (X 2 = 32.65, P < ). To determine the time at which adult male association peaks prior to known conception date, we fit a linear model (this time excluding random factors) and ran a piecewise regression to determine the most likely break point at which the relationship between day from conception and male association changes. At (SE ± 7.7) days prior to conception, the relationship changes from a positive slope (+0.017) to a negative, but nearly equivalent in magnitude, slope (-0.016), suggesting that there is an optimal 14

30 time for males to remain with females to ensure conception and paternity, and this drops off once pregnancy can be confirmed. Association with juvenile males, while consistently lower than adult males, showed a linear decline across the eight months from pre- to post-conception (Table I.2; Figures I.3a,b). Mother-son association patterns There was a steady decline in the proportion of mothers and sons that associate from weaning through adulthood (Figure I.4). Additionally, the predicted probability of a mother being seen with her weaned son, given his age, decreased nearly to zero by the time her son reaches age 20 (Figure I.5). Mother-son association post-weaning did not differ as expected from her association with unrelated males of comparable age. In fact, females associated more with their juvenile sons than they did with any given unrelated juvenile son (Z = -3.81, P < ). However, once their sons reached adulthood, mothers associated with adult sons just as much as unrelated adult males (Z = -0.16, P > 0.1). Despite these trends, 32% of mother-son pairs were never sighted together during the son s juvenile period, even though both were sighted frequently during that time. By contrast, 64% of mother-son pairs were never sighted together during the son s adult period, even though both were sighted frequently during that time. A more striking pattern emerges when taking into account the mother s reproductive state. When females were cycling, there was no difference in association with juvenile sons versus unrelated juvenile males (Z = , P = 0.122; Figure I.6). However, cycling females were rarely seen with adult sons compared to unrelated adult males (Z = , P < 0.05; Figure I.6). 15

31 Finally, group size was significantly larger when females were with their adult sons compared to when they were with their juvenile sons (Z = -4.94, df = 447, P < ). The predominant activity when females were with their adult sons was primarily traveling (40%, SE ± 9%) or resting (32%, SE ± 8%), and only rarely was the group socializing (13%, SE ± 6%). DISCUSSION In this study, we demonstrate how female-male associations vary by cycling state, relatedness, and age. Interestingly, the peak in adult female-adult male association occurred 14 days prior to the estimated conception date and association persisted briefly post-conception (Figure I.1). Some have hypothesized a 54-week gestation period, slightly longer than the 52- week rule of thumb (O Brien and Robeck, 2012), and our association patterns are consistent with that hypothesis. Interestingly, bottlenose dolphins are polyoestrous, going through 3-4 estrus cycles of approximately days each within a cycling season (Schroeder, 1990; Robeck et al., 2005; O Brien and Robeck, 2012). However, only one peak in above-average male association was detected within the four months prior to a known conception. It is unlikely that male interest would wane in between consecutive estrus cycles, even if anovulatory (e.g. elephants, Hildebrandt et al., 2011), because the cost of missing an ovulation would be high. Further, the resolution of our survey data would not likely be able to detect narrow, local association peaks prior to the term conception. 16

32 From pre- to post-conception, the average number of juvenile males present with adult females decreased significantly but to a small degree (Figure I.3a,b), while only adult males showed a sizeable peak in association at, or just prior to, conception (Figure I.1a,b). We suggest two, non-mutually exclusive hypotheses to explain this temporal pattern. First, we hypothesize that both juvenile and adult males can detect female pregnancy, and thus interest wanes following conception. While in some species young males may not yet be able to detect female fertility (e.g. leaf monkeys, Trachypithecus spp., Lu et al., 2012), our results are concordant with the hypothesis that males adjust their behavior after a female becomes pregnant. Some studies have recorded female fertility detection in dolphins by the presence of specific estrus behaviors based on captive dolphin (T. truncatus) observations (Muraco and Kuczaj, 2015) or pheromones released in urine (Muraco, 2015), but to date no study has documented the mechanism by which pregnancy is detected by conspecifics in bottlenose dolphins. Our second hypothesis to explain the temporal pattern of association between conceptive females and males is that females tolerate adult males only or elicit sexual advances towards adult males near ovulation. This may explain the higher than average adult male association prior to pregnancy, but consistently decreasing juvenile male presence prior to conception. Near conception, females associate with more adult males than juvenile males, until later in pregnancy when female-male association decreases in general. This suggests that juvenile males are not likely to be fathers, and contrasts work in other species. For example, young primate males have been known to engage in copulations (e.g. chimpanzees; Watts, 2015) or even father offspring (e.g. baboons; Alberts et al., 2003), albeit rarely. However, lone male dolphins cannot monopolize females, especially juveniles who have not yet formed a stable alliance. Further, established adult male alliances would easily steal a fertile female from a single 17

33 or even several cooperating juveniles. Across the board, it appears that females associate very little with juvenile males. We cannot differentiate who is driving the observed patterns so it is equally important to consider male trade-offs. Adult males are better equipped to outcompete juvenile males to gain access to fertile females because they are physically larger and have stable alliance partners, whereas juvenile males typically have not yet formed strong social bonds. Adult males may also be intolerant of juvenile males or even direct aggression towards them. A previous study found that male calves with more ties to adult males had reduced survival (Stanton and Mann, 2012), suggesting male competition during early stages of development. Given that adult males, and possibly juvenile males, can likely detect pregnancy as evidenced by reduced association with females post-conception, it is unlikely that postconception paternity confusion by multiple mating is a strategy utilized by Shark Bay dolphins. It is unclear whether infanticide is a driver in the evolution of dolphin behavior, but there are some documented cases of infanticide in dolphins (Tursiops spp.) at other locations (Patterson et al., 1998; Dunn et al., 2002; Kaplan et al., 2009; Robinson 2014; Perrtree et al., 2016). We argue that infanticide risk is not a driver in Shark Bay dolphin behavior, however, because individuals are philopatric, year-round residents where social bonds are important for fitness for both sexes (Krützen et al., 2004; Frère et al., 2010c; Stanton and Mann, 2012) and chances of paternity are relatively low because successful mating requires the cooperation of alliance partners (Krützen et al., 2004). Additionally, upon loss of offspring, it usually takes months before the female is fertile and the chance that a male has of fathering the next offspring might be small (Mann et al., 2000). 18

34 However, many females are consorted by multiple alliances during a single breeding year. The length of attractive periods is generally 7 days or less, though consortships can last several weeks (Connor et al., 1996) up to several months (Connor and Krützen, 2015), and mating probably occurs with more than one (if not all) alliance members during that time. So while paternity confusion via mating within and across alliances may occur during the preconception period, pregnancy probably can t be masked, as supported by the drop in male association upon conception. While some coerced consortships have been recorded during nonfertile periods (pre- and post-partum), these appeared to be an anomalous case restricted to one aggressive provisioned male alliance with almost unlimited access to food (Connor et al., 1996). Female choice is difficult to document in Shark Bay dolphins, where alliances harass females and coerce consortships during fertile periods. Mate choice is likely to be constrained in sexually coercive systems (e.g. chimpanzees; Muller et al., 2011). In Shark Bay dolphins, females are known to bolt to escape a male alliance, and it is fairly obvious that these females are trying to get away from aggressive males, as opposed to testing their physical vigour (Connor et al., 1996). An alternative explanation is that female tolerance of adult males is a mechanism of conflict management, seen in many primate species (Cords and Mann, 2014), as male aggression is known to be higher towards cycling females (Scott et al., 2005; Watson-Capps, 2005). In our study of female associations with juvenile and adult males near conception, all surveys with sons were excluded. We took a closer look at mother-son relationships to determine whether social mechanisms for close inbreeding avoidance were present (Clutton-Brock and Lukas, 2012), despite inbreeding values greater than expected by random mating in this population (Frère et al., 2010a). Mothers associate with their weaned sons less than their weaned daughters (Tsai and Mann, 2013), which echoes a more general pattern that females associate 19

35 less with males than with other females (Wallen et al., 2016). Interestingly, mothers associated with their weaned juvenile sons more than any given unrelated male of comparable age. We suggest that at young ages, males do not yet present a threat of inbreeding as they are still allocating energy towards growth, and haven t yet established stable alliances. Females can avoid mating attempts by sons, and male fertility is probably low in the juvenile period. Perhaps mothers are able to assist their sons socially while they are still at a young age, extending her maternal influence post-weaning (Gibson and Mann 2008a; b) either through information sharing, social support, or protection from older or more aggressive males (Stanton and Mann, 2012). Broken down by year, mothers gradually associated less with their weaned sons throughout from the subadult to adult period, and generally adult males only rarely associated with their mothers. However, association did not drop off to zero as expected. Only cycling females associated less with adult sons than with unrelated adult males. We cannot determine whether the mechanism is by mutual avoidance between mothers and adult sons, or whether competition between alliances for fertile females prevents sons from access to their mothers. Though mothers and their adult sons reduced association only during cycling periods, other nuanced behaviors may reflect attempts to minimize inbreeding between mothers and sons. Group sizes were larger when mothers were with adult sons compared to juvenile sons, suggesting a mechanism of social buffering. The social pressure of philopatry may mean that females only tolerate their sons after sexual maturity if there are many other individuals (and potentially her close associates) nearby. Further, groups were rarely socializing when mothers and sons were together, suggesting association is tolerated in non-mating contexts (e.g. large resting groups). To date, sons have not been observed mounting their mothers post-weaning (Mann, 2006). Given that dispersal is highly variable in Tursiops (e.g. no dispersal in Sarasota 20

36 Bay, FL: McHugh et al., 2011; dispersal (via genetic analyses) in NSW, Australia: Möller and Beheregaray, 2004), and that dolphins appear to tolerate some level of inbreeding (Frère et al., 2010b), sex-biased dispersal is likely explained by drivers other than inbreeding avoidance (Moore and Ali, 1984; Szulkin et al., 2013). We caution that survey association data introduces a potential gambit of the group bias, where individuals in the same survey may not necessarily be social partners. For example, if there are several adult females in a large group with several males, we can t be certain which female (if any) is driving male interest. To check this bias, we restricted our surveys to those where only one conceptive female was present, and our results stayed the same. In the future, direct interactions and association measures from individual focal follows may provide more complete picture on the causes and/or function of male-female associations across different reproductive states. Another caveat to this study is that kin association as measured is almost certainly an underestimate. While mother-dependent offspring relationships are known (Mann et al., 2000), there are likely many undetected kin relationships among adults that were dependent calves before the project started in Genetic sampling has enabled additional kin relationships to be discovered (Krützen et al., 2004), but there are likely some mother-offspring pairs that were analyzed as mother-unrelated male pairs (i.e. false negatives). Collection of physiological data via non-invasive methods has become increasingly important for understanding health and vulnerability of wild populations (Cooke et al., 2014). Techniques in reproductive endocrinology have allowed for the non-invasive study of physiology, most notably from urine (Preis et al., 2011) or fecal (Millspaugh and Washburn, 2004) metabolite analysis, though there are many unresolved issues (Goymann, 2012). In wild 21

37 cetacean populations, methods for detecting reproductive state are typically invasive. Blubber can be collected with a remote biopsy system, but is a crude measure of pregnancy, not the timing of ovulation (Perez et al., 2011), or tissue and sonograms can be collected from captured and/or temporarily restrained animals (Wells, 1991; Bergfelt et al., 2013). Fecal and urine samples can be collected from captive marine mammals (Robeck et al., 1993; Biancani et al., 2009; Amaral, 2010), and more recently, scat collection is possible with wild cetaceans using detection dogs (e.g. right whales, Eubalaena glacialis; Rolland et al. 2005; killer whales, Ayres et al., 2012). Non-invasive detection of reproductive state in smaller delphinids remains a challenge. For human observers without easy access to physiological indicators, behavior and association patterns may serve as accurate indicators of female reproductive state in wild populations. In the case of bottlenose dolphins, fertile periods can be determined post-hoc based on pregnancy and birth of a calf, but might also be obvious based on male attention, behavior and association. Mounting behaviors, while common amongst immature dolphins, are rarely observed between adults of the opposite sex (Mann, 2006), but probably correspond to periods of oestrus. Lactation is conspicuous in dolphins because the presence of a dependent calf in infant position indicates nursing offspring (Mann et al., 2000) and mammary glands are visibly swollen up until the time of weaning. However, bottlenose dolphins have overlapping reproductive states, and become attractive to males prior to weaning dependent offspring (Mann et al., 2000). Pregnancy cannot be easily detected until the later stages based on female girth (J.M., personal observation), unless the animal is captured and restrained for blood testing or ulrasonographic analysis (Wells and Scott, 1990; Wells et al., 2004). If there is foetal loss, especially early foetal loss, then the pregnancy and corresponding fertile periods, might not be known. 22

38 Association patterns as demonstrated in this study may be a useful tool to model known conceptions, at least in species with predictable temporal sexual segregation such as the bottlenose dolphin. Empirical association records may be useful for informing computational models to predict reproductive events that go undetected via observational methods, such as pregnancies lost in utero, or ovulations that were not followed by a conception. Future work should investigate the predictive power of intersexual associations with known pregnancies, as a model for determining missed ovulations or conceptions. 23

39 FIGURES AND TABLES Figure I.1a. The fitted quadratic model with 95% confidence intervals for the predicted number of adult males associating with females by day from conception. The solid vertical black line at 0 indicates predicted conception date. The dashed red line indicates the breakpoint at which adult male association starts to decline. 24

40 Figure I.1b. The fitted quadratic model with raw data points and 95% confidence intervals for the predicted number of adult males associating with females by day from conception. The solid vertical black line at 0 indicates predicted conception date. The dashed red line indicates the breakpoint at which adult male association starts to decline. 25

41 Figure I.2. Box and whisker plot showing the number of adult males associating with females by day from conception. The solid vertical black line at 0 indicates predicted conception date. The dashed red line indicates the breakpoint at which adult male association starts to decline. 26

42 Figure I.3a. The fitted linear model with 95% confidence intervals for the predicted number of juvenile males associating with females by day from conception. The solid vertical black line at 0 indicates predicted conception date. The dashed red line indicates the breakpoint at which adult male association starts to decline. 27

43 Figure I.3b. The fitted linear model with raw data points and 95% confidence intervals for the predicted number of juvenile males associating with females by day from conception. The solid vertical black line at 0 indicates predicted conception date. The dashed red line indicates the breakpoint at which adult male association starts to decline. 28

44 Figure I.4. Proportion of mother-son pairs seen together by son s age in years postweaning. Sample sizes are above bars. 29

45 Figure I.5. Predicted probability with 95% confidence interval of mother-son pairs being sighted together post-weaning, given son s age. 30

46 Proportion time together when female cycling * Adult NS Juvenile Figure I.6. Proportion of time that females were with adult or juvenile males when she was cycling. Black bars indicate sons, white bars indicate unrelated males. 31

47 Table I.1. Parameter estimates for fixed effects in the quadratic regression model for adult male association. Estimate ± SE t P R 2 Intercept ± < Day from conception ± (Day from conception)^ ± <0.0001*** Table I.2. Parameter estimates for fixed effects in the linear regression model for juvenile male association. Estimate ± SE t P R 2 Intercept ± < Day from conception E04 ± 3.389E ** 32

48 CHAPTER II The ecological costs to females in a system with allied sexual coercion 2 INTRODUCTION Sexual coercion, an extreme example of sexual conflict, is defined as when males, at some cost to females, direct force or the threat of force toward females to increase their chances of mating when females are fertile, and to decrease females chances of mating with other males (Smuts and Smuts 1993). Coercion, an adaptive male strategy to overcome female resistance to mating and monopolise breeding opportunities, is an important force in sexual selection (Clutton-Brock and Parker 1995) and can potentially even lead to divergence and speciation (Panhuis et al. 2001). Furthermore, direct costs to females due to male sexual aggression can be severe and include injury (Le Boeuf and Mesnick 1991; Hiruki et al. 1993), increased energy expenditure (Watson et al. 1998), increased mortality (Réale et al. 1996), physiological stress (Muller et al. 2007), and decreased reproductive success (Hiruki et al. 1993; Ojanguren and Magurran 2007; den Hollander and Gwynne 2009; Gay et al. 2009; Rossi et al. 2010; Takahashi and Watanabe 2010). Nonetheless, because documenting such fitness costs is challenging, particularly in wild, long-lived animals, some researchers have examined the behavioural and ecological costs females experience as a result of male coercion, which may or may not have consequences for fitness. For example, studies documenting changes in movement 2 See Appendix A. Citation: Wallen, M. M., Patterson, E. M., Krzyszczyk, E. & Mann, J. (2016). The ecological costs to females in a system with allied sexual coercion. Animal Behaviour, 115,

49 and ranging (e.g. Grevy s zebra Equus grevyi, Sundaresan et al. 2007), activity patterns (e.g. southern elephant seals Mirounga leonina, Galimberti et al. 2000; mollies Poecilia spp., Heubel and Plath 2008; humpback whales Megaptera novaengliae, Cartwright and Sullivan 2009; and guppies Poecilia mexicana, Köhler et al. 2011), and sociality (e.g. guppies Poecilia reticulata, Darden et al 2009; Darden and Watts 2012) suggest that male coercion influences important aspects of female behavioural ecology and likely fitness. Yet among these studies, few have examined the impact males have on female behaviour or fitness when they act collectively, i.e., coalitionary or allied aggression, perhaps because outside of humans (Rodseth and Novak 2009), some non-human primates (chimpanzees Pan troglodytes: Watts 1998, Muller et al. 2009, Connor and Vollmer 2009; baboons Papio: Noë 1992; spider monkeys Ateles: Link et al. 2009), and some bottlenose dolphin (Tursiops spp.) populations (Connor et al. 1992a, Connor and Vollmer 2009), allied males rarely direct aggression towards females. In several long-term studies of bottlenose dolphins, researchers have documented a sexually coercive mating system in which adult males form long-term, stable alliances (Connor and Vollmer 2009) of variable size (Wells 1991; Connor et al. 2001; Owen et al 2002; Wiszniewski et al. 2012) that cooperate to consort and mate with individual, primarily cycling, females (Connor et al. 1992a; Smolker et al. 1992; Connor et al. 1996). Consortships are typically initiated by aggressive herding behaviours such as biting, hitting, chasing, and threat displays or captures, followed by intermittent aggression throughout the consortship (Connor et al. 1992a; Connor and Smolker 1996). Among Indian Ocean bottlenose dolphins (Tursiops cf. aduncus) in Shark Bay, Australia, preliminary evidence suggests that allied males influence female ecology. Previous work found that females spend more time in deeper water and less time in shallow water when in consortships (Watson-Capps 2005). Although the benefits or costs of 34

50 this shift are not fully understood, changes in depth use suggest that male coercion may impact female spatial ecology. Shark Bay dolphin spatial ecology has been previously described in some detail. Individuals exhibit bisexual philopatry and have large, overlapping home ranges that are stable through time (Tsai and Mann 2013). Habitat use is influenced by both predator (tiger shark, Galeocerdo cuvier) and prey distributions on large and small spatial scales (Heithaus and Dill 2002; 2006), meaning even small shifts in space use could have potentially serious ecological outcomes for dolphins. However, such shifts likely have the greatest impact on female ecology given that females exhibit habitat specific foraging specialisations (Sargeant et al. 2007, Mann and Sargeant 2003; Sargeant et al. 2005; Mann et al. 2008; 2012), have smaller home ranges, and lower habitat use diversity compared to males (Patterson 2012). For example, some females specialise in a foraging tactic known as sponging, which involves the use of marine sponges as tools and only occurs in the deep channels where sponges and appropriate prey are found (Mann et al. 2008; Sargeant et al. 2007; Patterson and Mann 2011). Sponger females could be severely impacted if consorting males move them away from the channel habitat. In contrast, individual males and alliances have much larger home ranges and greater habitat use diversity (Randić et al. 2012; Patterson 2012), which likely relates to their need to roam the bay to find and maintain access to fertile females. Thus, not only is efficient space use inherently important for survival, but the observed variation among individuals and among sexes is an explicit representation of individual ecological needs. When considering that male and female space use must coalesce during consortships, three scenarios are possible. First, it may be that males spatially sequester females by consorting with them in accordance with their alliance s space use. Here one would expect substantial 35

51 ecological costs to females, and no such costs to males. Second, it may be that males spatially sequester females to some extent, but also partially adjust their alliance s space use to temporarily match that of fertile females. Here one would expect ecological costs to both sexes, the magnitude of which would depend on the relative space use shifts for each sex. Finally, it may be that males do not spatially sequester females and instead temporarily adjust their alliance s space use to match that of their targeted mate's range, i.e., males go where the fertile females are and follow them around. Here one would expect males, but not females, to suffer an ecological cost. Given the aggressive nature of consortships, the first or second scenario, both of which impose some ecological costs on females, seem most likely. Thus, we hypothesise that males present an ecological cost to adult females by altering female space use, specifically, their ranging and habitat use. If alliances sequester females to their own, much larger home ranges, females will likely be far from their core home range area and their preferred foraging habitats. Accordingly, we predict that (1) females will be farther from their home range core (i.e. the centroid) when they are with more than one male compared to when they are not, and that (2) females will use their preferred habitat less when they are with more than one male compared to when they are not. However, this does not preclude consortships from affecting male space use. Nonetheless, given that males have larger home ranges and greater habitat use diversity than females, even if males do alter their space use during consortships, we expect the relative impact of consortships on spatial ecology to be greater for females than for males. Thus, we predict that (3) if males do experience space use shifts during consortships, such shifts in both ranging and habitat use will be relatively greater for females than for males when the sexes are together. 36

52 Female-biased space use shifts would suggest that females suffer an ecological cost in this coercive mating system. METHODS Study population and site Our study population consists of individually recognised wild Indian Ocean bottlenose dolphins (Tursiops cf. aduncus) residential to Shark Bay, Western Australia (Mann et al. 2000; Tsai and Mann 2013). As part of the Shark Bay Dolphin Research Project (SBDRP), researchers have collected behavioural, demographic, reproductive, ecological, social, and genetic data on >1800 dolphins since Individuals are distinguished using standard dorsal fin identification techniques (Würsig and Würsig 1977). Sex is determined by the presence of a dependent calf, views of the genital area (Smolker et al. 1992), and in a few cases, DNA (Krützen et al. 2004). Age is determined from known or estimated birthdates (if seen as a calf), physical and behavioural characteristics (Mann and Smuts 1999), and/or the presence and degree of ventral speckling (Krzyszczyk and Mann 2012). Our main study site is a 300 km 2 area of the eastern gulf of Shark Bay (25 47 S, E) within a UNESCO World Heritage Site, and as a result, remains relatively pristine with low human impact. Habitat in the study area, as defined by Patterson (2012), consists of six distinct types (average depths reported relative to datum): channel (7.13 m) with a substrate of rock, shell, and coral debris, deep open (6.56 m) with a mixed sand, silt and clay substrate, seagrass 37

53 beds (2.00 m) with continuous seagrass coverage (predominantly Amphibolis antarctica and more sparsely Posidonia australis), sand flats (-0.11 m) with continuous sand coverage, and two edge habitats: deep ecotone, the transition zone between a shallow (seagrass beds or sand flats) and deep (channel or deep open) habitat and shallow ecotone, the transition zone between two shallow habitats (seagrass beds and sand flats). Relative habitat availability was calculated as the proportion of the study area with coverage of that habitat type (Patterson 2012). Data collection Survey records Data collection for the SBDRP consists primarily of observational, boat-based records. For this study data were drawn from surveys, which are opportunistic sightings of dolphins, conducted from A survey began when observers are close enough to identify or photograph individuals. Scan sampling (Altmann 1974; Mann 1999) is used for the first five minutes of a survey to determine group composition and activity. Group membership was determined using a 10-meter chain rule, where all individuals within 10 meters of another group member were considered to be in association (Smolker et al. 1992). Predominant group activity was recorded as the behaviour that 50% of the group members engaged in during the scan in the first five minutes. Spatial data (latitude and longitude) were also collected during each survey, to later be used in classifying each survey s habitat based on the aforementioned habitat classifications. If an individual had more than one survey sighting in a day, we restricted our 38

54 dataset to include only the last survey point per individual per day to reduce spatial and temporal correlation. Female cycling Male interest in females is expected to vary with female reproductive status. Thus, we included female cycling status (whether or not she was cycling) as the first factor in our study. Bottlenose dolphins are seasonally polyoestrous with an estimated 6-month peak cycling period prior to the ~12-month pregnancy (Schroeder 1990; O Brien and Robeck 2012). Accordingly, we assigned cycling periods as the 6-month period before the start of known pregnancies (from known births). However, this definition biases observations against females who have produced no or few surviving calves, or have long intervals between nursing calves since they are potentially fertile during periods when no pregnancies are confirmed by a birth. Consequently, we included additional time periods in which a female could be cycling, based on the period: a) from the female s 12 th birthday (when 42.4% of first births occur, while 22.0% occur at age 11, N = 59 adult females where age of first birth was known) until her first known pregnancy, and b) one year prior to weaning date for calves that nursed beyond the age of four, the average weaning age (average interbirth interval between surviving calves is 4.7 years; Mann et al. 2000). This method overestimates cycling periods (increasing the risk of Type II error), but since females go through several non-conceptive cycles per year, it conservatively accounts for potential fertile periods. Since most of our data (76% of surveys) were collected during the 6 months prior to the November peak births/conceptions (Mann et al. 2000), we have adequate coverage over periods of cycling. Females were considered to be non-cycling if they were (1) 39

55 pregnant, or (2) nursing a calf that was at least 12 months from its weaning date. Thus, our cycling factor has two levels: cycling and not cycling. Females under age 12 were excluded from analysis. Male presence Individual pairs or trios of males (first-order alliances) typically have very strong bonds, with half-weight coefficients of association (COAs) as high as (Connor et al. 1992a, b; 2011). Second-order alliances are composed of two or more first-order alliances (Connor and Vollmer 2009) with COAs ranging from (Connor & Krützen 2015), and third-order alliances are composed of second-order alliances and their close male associates with slightly lower COAs ranging from (Connor et al. 2011). Because males are found so frequently with their alliance partners and very rarely with non-partners, as a proxy for consortships we used the presence of more than one adult male in a survey. Males were considered to be adult at 12 years of age or older (Cheal and Gales 1992). Thus, we used male presence as the second factor in our study, having two levels: >1 male present, in which more than one adult male was present in the survey (representing the minimal unit for allied coercion), and males absent, in which no adult males were present in the survey. Here, single-male associations were omitted. Males and females are monomorphic and there is no evidence that a single male can monopolise a female. Although allied aggression towards females is frequently observed (Connor et al. 1996; Scott et al. 2005), aggression by a lone adult male towards an adult female has never been observed in 28 years of study and 3467 hours of focal observation on 127 adult females (Scott et al. 2005; unpublished data). Further, associations between single males and single or multiple 40

56 females are rare (Table II.1). Nevertheless, for a subset of females in our sample, we included a third level for the male factor: surveys in which a single male was present with at least one adult female (see Supplementary Material). Data analysis Cycling status (cycling or not cycling) and adult male presence (males absent, one male present, >1 male present) were included as factors in our analyses. Since reproductively successful females spend most of their adult life pregnant and/or lactating, not cycling, and rarely in association with males (Gibson and Mann 2008), we considered the baseline state to be surveys when females were not cycling and males were absent. We conducted a similar analysis on males to examine if their behaviour is impacted by the presence of a female. Males spend most of their time with each other (Connor et al. 1992a, b; 2011) and not in consortships, so we considered the male baseline state to be surveys when no adult females were present (females absent) and there was at least one additional adult male present. The male baseline was compared to surveys where males were with at least one female (cycling or non-cycling), resulting in a single factor with three levels for the male analysis (no females, only non-cycling female(s) present, and only cycling female(s) present). Dolphins were included in the sample if they had a minimum of six survey sightings on different days in each factor level (e.g. cycling female with males; one point possible for each habitat based on habitat availability), and the same females (N = 32) and males (N = 73) were used for all analyses (the smaller sample size for females reflects the additional cycling restriction). A subset of 11 females with adequate survey data was 41

57 analysed using the single-male comparison (results shown in Appendix A). For all analyses regardless of overall significance tests, we conducted a priori contrasts between all three levels of male presence. Ranging As a measure of core ranging we calculated a baseline centroid for each individual by averaging all GPS locations from survey records in the baseline state. Centroids for males and females were calculated for their baseline state excluding sightings from all other factor levels. To examine if ranging differed with cycling status and male presence (for females) or with female presence including cycling status (for males), we compared the distances between the baseline centroid and each survey GPS record within each factor level using permutation tests with dolphin ID as a blocking factor. While somewhat counterintuitive, we also compared the distance of the baseline centroid to each baseline survey point used in its calculation, as this serves as the best measure of expected distance from the baseline centroid if there were no effect of male presence or cycling status. We then compared the distances from the baseline for each sex when they were found together to determine if females and males differentially shifted in ranging. The average number of surveys (± SD) per female, per factor level was ± (not cycling, males absent), ± (not cycling, >1 male present), ± (cycling, males absent), and ± (cycling, >1 male present). The average number of surveys (± SD) per male, per factor level was ± (females absent), ± (cycling female(s) present) and ± 8.70 (only non-cycling female(s) present). 42

58 To further understand how cycling and male presence influence female ranging, and how female presence influences male ranging, we also calculated the average pairwise distance between all sightings for each individual within each factor level (hereafter average ranging distance). We examined the effect of cycling, male presence, and their interaction (for females), and the effect of female presence (for males) on this average ranging distance using permutation tests. Habitat use After spatially intersecting our survey sightings with the six habitat classifications, we quantified individual preferred (hereafter primary) habitat use using selection ratios. Selection ratios indicate habitat selection by an individual while correcting for habitat availability. Selection ratios were calculated as the proportional use of a habitat divided by the proportional availability of that habitat (Manly et al. 2002). Values greater than one indicate selection above expected based on availability, values less than one indicate selection below expected based on availability, and values equal to one indicate selection in accordance with availability. We ranked each individual s baseline habitats (1, 2, 3, etc.) in descending order by their selection ratio, with the primary (1 ) habitat having the highest value. Habitat use shifts were examined as a function of cycling status and male presence (for females) and female presence including cycling status (for males), using distanced-based, permutation MANOVAs with dolphin ID as a blocking factor. In this analysis the last-ranked habitat (6 ) was dropped since otherwise the response would be linearly dependent. 43

59 Because many dolphins, mostly females, specialise in habitat-specific foraging tactics (Sargeant et al. 2007, 2005; Mann et al. 2008; Patterson and Mann 2011), we examined primary habitat use alone with permutation independence tests. We expected the primary habitat to be the most important habitat for females and likely the most affected by male presence. We then compared the primary habitat use selection ratio for each sex when they were found together to determine if females and males differentially changed use of their primary habitat. The average number of surveys (± SD) per female, per factor level was ± (not cycling, males absent), ± (not cycling, >1 male present), ± (cycling, males absent), and ± (cycling, >1 male present). The average number of surveys (± SD) per male, per factor level was ± (females absent), ± (cycling female(s) present), and ± 8.25 (only non-cycling female(s) present). See the Interactive Map for a visualization of habitats in the study area as well as representative male and female spatial data by factor level. Statistical considerations All of our response variables deviated significantly from normality (all Shapiro-Wilk tests P < 0.01), thus permutation tests were used in all analyses (Anderson 2001). Permutations (1000 randomizations) were performed using the coin package (Hothorn et al. 2006) in R, version (R Core Team 2015), except for the female habitat and ranging distance analyses in which custom permutations tests were written using the lme4 package for R (Bates 2014) to test for interaction (Anderson 2001). Permutation MANOVAs were performed using the adonis() function in the vegan package for R (Oksanen et al. 2013), which permutes the raw data and calculates a pseudo-f using a multivariate distance measure, here the Bray-Curtis distance. For 44

60 all analyses, the interactions were tested and tests for main effects were only performed if interactions were determined to be non-significant. Significance was set at α = 0.05 for all tests, and Bonferroni adjustments were applied where there were multiple contrasts and the corrected P-values are reported. Female centroid distances were natural-log transformed to correct for heteroscedasticity (Levene s test P < 0.01), but raw distance values are reported for illustrative purposes. RESULTS Female associations with males Table II.1 shows that adult females spend most of their time either alone or with other adult females. Females are with >1 male about one-third of the time when cycling, and are very rarely with lone males regardless of reproductive state. These descriptive data provide context for our results. For all the following analyses, results from the single-male associations were either not significant or largely intermediate between male-absent and multi-male associations (see Appendix A, Figure S1a-c, Table S1). As evident from Table II.1, the instances of single-male association with females are rare and our sample sizes were limited. See the Supplementary Material in Appendix A for all single male results and discussion. 45

61 Ranging There was no interactive effect of cycling status and male presence on distance from baseline for females (F = 0.24, P = 0.63), so main effects are reported. Both cycling status (Z = , P < ) and male presence (Z = -7.08, P < ) influenced the distance (natural logtransformed) a female was found from her baseline centroid. Females were sighted farthest from their baseline centroid when they were with males, and also when they were cycling (Fig. II.1a). In contrast, male distances from their baseline centroid were unrelated to female presence or female cycling status (maxt = 1.63, P = 0.20; Fig. II.2a). In comparing male and female distances from their baselines to when there was at least one female and >1 adult male in the group, the distances of each sex to their respective baseline centroids differed (Z = -2.35, P = 0.02), with males being farther from their baseline centroid than females. Females average ranging distance was not influenced by an interaction between cycling status and male presence (F = 0.13, P = 0.72) so main effects are reported. Both cycling status (Z = -2.55, P = 0.009) and male presence (Z = -3.06, P = 0.003) affected a female s average ranging distance (Fig. II.1b). Females had the highest average ranging distance when cycling and >1 male was present, which was significantly higher than the baseline state only (Z = -3.56, P < ) after a Bonferroni correction (Fig. II.1b). However, cycling females without males also had a higher average ranging distance than females in the baseline state (Z = -2.85, P = 0.001; Fig. II.1b). Males overall average ranging distance was influenced by female presence (maxt = 5.09, P < ). Males had the highest average ranging distance when no females were present compared to when non-cycling female(s) (Z = 4.79, P < ) and cycling females (Z = -2.77, 46

62 P = 0.005) were present (Fig. II.2b). Average male ranging distance was greater when cycling females were present compared to only non-cycling female presence (Z = 3.24, P = 0.002; Fig. II.2b). Habitat use The interaction between male presence and cycling status on overall female habitat use was not significant (Table II.2). However, overall female habitat use did change significantly as a function of male presence, but not cycling status (Table II.2). Per our a priori prediction, we examined the effect of cycling and male presence on female primary habitat alone (Fig. II.1c). There was no interaction between cycling status and male presence on primary habitat use (F = 0.44, P = 0.53) so only main effects are reported. Consistent with the results from the MANOVA, all females used their primary habitat less when in the presence of >1 male (Z = 3.32, P = ), independent of cycling status, which itself was not significant (Z = 0.28, P = 0.78). In the baseline state, females were sighted in their primary habitat 62% of the time, and this decreased by an average of 15% when >1 male was present. Overall male habitat use changed as a function of female presence (Table II.2), as did the primary habitat when analysed alone (maxt = 3.54, P = 0.001). Males used their primary habitat less when either cycling or non-cycling females were present compared to when no females were present (cycling females: Z = -4.25, P < ; non-cycling females: Z = 2.59, P = 0.008; Fig. II.2c). However, male habitat use did not differ depending on if they were with cycling females or non-cycling females (Z = 0.23, P = 0.84). In baseline, males were sighted in their primary habitat 54% of the time, and this decreased by an average of 4% when either cycling or non- 47

63 cycling females were present. In comparing the reduction in primary habitat use between males and females from their baseline to when there was at least one female and >1 adult male in the group, the reduction in primary habitat use did not differ between the sexes (Z = 0.54, P = 0.58). DISCUSSION Our findings are consistent with the sexual coercion hypothesis that allied males impose ecological costs on females, in that females experienced space use shifts when in the presence of more than one adult male. Specifically, females were sighted farthest from their baseline centroid, had greater average ranging distances, and showed altered habitat use when with more than one male, regardless of cycling status. Interestingly, cycling status alone also affected female ranging in that females had greater ranging distances and were found farther from their baseline when cycling, regardless of male presence. This suggests an overall change in ranging, although not habitat use, during cycling that may be unrelated to mating. Together, our results suggest that allied male coercion imposes ecological costs on female bottlenose dolphins, but some range shifts may occur outside of consortships when females are cycling. Males did not experience as great a shift in their spatial ecology during mating contexts as did females. Males were found equally far from their baseline whether or not a female (either cycling or non-cycling) was present. However, male average ranging distance decreased when with cycling and non-cycling females. This is unsurprising as males are expected to range widely when searching for females, and preliminary data suggest males have smaller home ranges during peak breeding in the spring (Randić 2008). When a female was present, male average 48

64 ranging distance was higher if the female was cycling, which suggests males a) match movement of the females they are consorting, and/or b) maintain greater ranging to avoid detection or conflict with other alliances. Both predictions corroborate our finding that female average ranging distance was higher when cycling. While males did alter their overall habitat use and decreased their primary habitat use in the presence of females, this did not depend on a female s cycling status. The lack of a cycling status effect here may indicate that males have imperfect detection of female reproductive states. Females nurse calves for 2.5 to 9 years and often wean their calf during the next pregnancy (Mann et al. 2000; unpublished data). With such long and variable lactation periods, male ability to detect or predict female fertility is likely imperfect. Males have also been observed to consort non-cycling and pregnant females (Connor et al. 1996, Furuichi et al. 2014), which may be a male bonding tactic rather than about increasing the likelihood of conceptions per se. When directly comparing centroid and habitat shifts during sightings of possible mating contexts (i.e. > 1 male and 1 female together in surveys), we found that males experienced larger ranging shifts than females, and there was no habitat shift sex difference. However, while the presence of males altered females normal ranging patterns (Fig. II.1a), the same was not true for males when in the presence of females (Fig. II.2a); male centroid distance remained constant regardless of female presence (cycling or non-cycling). Males typically have larger home ranges than females (Owen et al. 2002; Urian et al. 2009; Patterson 2012; Randić et al. 2012; Sprogis et al. 2016; but see Tsai and Mann 2013), so any shift they experience likely has less of an impact on their overall space use compared to a similar shift in females. In fact, increased ranging may actually be beneficial for males since the size of their home ranges is associated with their coercive mating tactics (Randić et al. 2012), and alliances tend to range more widely to 49

65 maximise their search area and access to females (Randić 2012). Further, while both males and females experienced similar primary habitat use shifts in mating contexts, adult males have higher habitat use diversity than adult females (Patterson 2012), and don t specialise in foraging tactics nearly as much as females (e.g. Mann and Sargeant 2003; Sargeant et al. 2005; Mann et al. 2008; 2012), so a shift in primary habitat likely has less of an impact on male foraging than it would for more specialised females. While males experienced smaller shifts in their spatial ecology when with females, our results clearly demonstrate ranging shifts for females in consortship contexts that likely impose ecological costs. For example, adult female home ranges in Shark Bay average 51 km 2 (Tsai and Mann 2013), compared to 76 km 2 for adult males (Randić et al. 2012). Thus, the average centroid shift of 3 km when cycling and with males reported here would likely place a female near the edge or outside of her home range, if, as a conservative example, a female has a circular home range with an area of 51 km 2 and a radius of 4 km. Interestingly, female sightings were farther apart (ranging distance) and farther from their baseline centroid when they were cycling regardless of male presence. Thus, in general, cycling females appear to increase their ranging. For some of these females that still have dependent calves, additional ranging may be especially costly, which raises the question of why cycling females might increase their ranging behaviour. We propose three possible explanations for increased movement during cycling, all of which warrant further investigation. First, cycling females might increase ranging to avoid detection by allied males. Second, females might increase ranging to avoid consortships with specific alliances (e.g. reduce probability of incestuous matings, see Frère et al. 2010a) and/or increase the chance of encountering preferred alliances or multiple male partners. Third, as an artifact of our study methods, females might be 50

66 sighted between consortships after males have already moved them far from their core ranging area. Increased traveling during cycling has been reported in other systems: in Mahale chimpanzees (Pan troglodytes schweinfurthii), estrous females travelled greater distances (Hasegawa 1990) and spent more time moving than their anestrous counterparts (Matsumoto- Oda and Oda 1998). In giraffes (Giraffa camelopardalis), cycling females tended to locomote more than pregnant females (del Castillo et al. 2005). Lactating Grevy s zebra have elevated speeds compared to non-lactating females, and these females experience the most male harassment (Sundaresan et al. 2007). In black-faced impalas (Aepyceros melampus petersi), postparturient females foraged in different microhabitats and elevations than pre-parturient females (Matson et al. 2007), suggesting finer scale habitat decisions may indeed be influenced by reproductive status alone. In contrast, in our study female cycling status influenced average ranging distance and distance from baseline centroid, but not habitat use. This suggests that although females do increase ranging and possibly their home range, they appear to do so in way that maintains their preferred habitat use. Such range shifts are not likely as costly as those experienced in a consortship context. Females experienced overall habitat use shifts, and importantly a reduction in primary habitat use, when with more than one male. Reduced primary habitat use has direct implications for female foraging behaviour since many females specialise in habitat-specific foraging tactics (Sargeant et al. 2007, 2005; Mann et al. 2008; Patterson and Mann 2011). For instance, given that spongers are almost never sighted greater than 6 km from their channel habitat (Mann and Patterson 2013), consorting by males almost certainly takes them in to unfamiliar habitat outside of their baseline area. In fact, one sponger female in this study reduced her channel habitat use by 36% when she was with males. In such a scenario, habitat and site unfamiliarity may present 51

67 significant challenges, such as extra search effort to find sponge tools and/or prey, which may alter foraging success (Patterson et al. 2015). Indeed, previous work found that females foraged less during consortships (Watson-Capps 2005), which could be because they were spending less time in their primary foraging habitat. An examination of group behaviour data from the surveys used in this study indicates that females have reduced foraging budget in consortship contexts. Foraging was the predominant activity of groups consisting of a female and more than one male only 20% of the time, compared to 32% for groups of females without males. While such a reduction suggests an energy intake cost to females as a result of male coercion, we emphasise that these group behaviour data are not ideal for examining individual female changes in activity budgets (Karniski et al. 2014). Instead, individual focal follow data should provide a more accurate picture of female activity budgets with respect to cycling status and male presence, a topic of our future work. Together, our results suggest that males spatially sequester females during consortships, from which they may gain several benefits. By sequestering females, males likely impact the ability of females to counter their efforts. For example, on a number of occasions, we have witnessed adult females appearing to assist other females in consortships (see also Connor et al. 2006). In these cases, the assisting female may join a group where males are harassing a close associate, and remain in close proximity or even establish physical contact with her. While males may try to separate these females, the females sometimes succeed in leaving the group. By spatially moving females away from their close associates, males might reduce the risk that they will receive aid. Indeed, female social relationships are closely tied to their home ranges (Frère et al. 2010b; Mann et al. 2012), so females displaced from their preferred areas may not have their common female associates available, albeit temporarily. 52

68 Another benefit of sequestration is that males may be able to reduce competition from other alliances. Alliances vary in size (Connor et al. 1999; Randić et al. 2012) and can be fiercely competitive, with larger alliances often defeating smaller ones (Connor et al. 1999). Presumably this is why males cooperate in the first place to increase their ability to maintain exclusive access to a female even though they must share in matings with alliance members. The best tactic for males might be to remove a female from an area where other males are likely to search. Finally, by using coercion within or outside of female cycling periods, males may be acting to increase their own reproductive success, as has been found in chimpanzees (Feldblum et al. 2014). While our data indicate that male coercion negatively impacts female spatial ecology, three alternative, but non-mutually exclusive, explanations are possible. First, females might actually alter their space use when cycling, and males simply follow them in order to gain mating access. As noted above, females of a variety of mammalian species increase their ranging during cycling for many reasons. However, we failed to find an overall effect of cycling on habitat use. Thus for this alternative hypothesis to hold, females would need to specifically alter their preferred habitat use only in the company of males, or, males would need to only follow a female if she is both cycling and in a non-preferred habitat. While this is certainly possible, we are unaware of any data that support this somewhat unusual scenario. Second, females might shift their space use to avoid males, and thus the ecological shifts represent the costs of a female counterstrategy to male aggression, rather than the cost of coercion itself. While we cannot rule out this possibility, in such a scenario the threat of male coercion would still impose an indirect cost on females if avoidance requires moving away from preferred areas. Finally, rather than experiencing direct costs from sexual coercion, females may be experiencing indirect costs as a 53

69 result of a male strategy to reduce intrasexual competition via spatial segregation. Since these two explanations result in the same behaviour (females mating with sequestering males, and not other competing males), we are unable to disentangle the root source of these ecological costs. If alliance formation and maintenance is actually a strategy to reduce male competition, particularly in areas where males have a high encounter rate (Connor and Whitehead 2005; Connor and Krützen, 2015) and encounters with cycling females are rare, then coalitionary aggression, sequestration, and monopolization of individual females may be the best way for males to achieve mating success. Future studies employing focal follow methods on individual females with and without males, in all reproductive states, may prove useful in ruling out these alternative explanations. Currently, the costs of sexual conflict are not well understood, especially among longlived mammals where fitness outcomes can take decades to assess (Aloise King et al. 2013), and female counterstrategies may be in place (Palombit 2014). However, our results suggest that in bottlenose dolphins, males have a significant impact on female space use that likely impacts foraging behaviour and potentially fitness. Given that female bottlenose dolphins spend weeks or even months in consortships (Connor et al. 1996), these spatial, ecological costs may be long lasting. For reproductively successful females that nurse calves for three or more years and cycle about every 4 5 years, consortship frequency, and thus the costs of male coercion may be low. However, for females with low calving success, repeated annual consortship events might have more severe negative impacts. As such, the fitness costs of sexual conflict to females in the bottlenose dolphin mating system deserve further study. Our study contributes to the growing body of literature documenting behavioural costs to coercive mating, and provides clear evidence for the costs of allied male aggression in a non- 54

70 human species. In other coercive systems, the threat of male coercion forces female avoidance (e.g. guppies Poecilia reticulata, Darden and Croft 2008), and even sexual social segregation (e.g. small-spotted catsharks Scyliorhinus canicula, Wearmouth et al. 2012). Male chimpanzees lead females away in consortships, effectively mate guarding (Tutin 1979). However, all known cases involve single male harassment of females (but see Watts 1998), and we are unaware of any cases where males engage in stable and long-term cooperation to coerce females as in bottlenose dolphins. Importantly, few studies have looked at the relationship between coercion, female reproductive status, and the effects on individual ecology. Such an integrative approach is important for understanding the mechanisms by which coercion can influence female fitness, and ultimately the evolution of mating systems. 55

71 FIGURES AND TABLES Distance from baseline centroid (km) (a) a b Male presence 0 males >1 male c d 5 (b) b Average ranging distance (km) a ab b 0 3 (c) Primary habitat selection ratio 2 1 a b a b 0 Not cycling Cycling 56

72 Figure II.1. Adult female (N = 32) (a) distance from the baseline centroid, (b) average ranging distance, and (c) primary habitat selection ratios for each combination of cycling and male presence factor levels. Error bars indicate ±1 SE. Letters indicate significant differences after Bonferroni adjustments. Distances in (a) were natural log-transformed for analysis but raw distances are shown for illustrative purposes. 57

73 Distance from baseline centroid (km) 4 (a) Average ranging distance (km) (b) a b c 0 3 (c) Primary habitat selection ratio 2 1 a b b 0 Non cycling Cycling No females female(s) female(s) 58

74 Figure II.2. Adult male (N = 73) (a) distance from the baseline centroid, (b) average ranging distance, and (c) primary habitat selection ratios for each factor level. Error bars indicate ±1 SE. For distance from baseline centroid, no significant differences across groups. For average ranging distance and habitat selection ratios (b), letters indicate significant differences after Bonferroni adjustments. Table II.1. Proportion of sightings per individual per factor level used in the full data set. (32 females; N = 4611 surveys and 7004 female sightings). For analysis, surveys with one female and one male were lumped with those that had more than one female and one male to form the one-male factor level. SIGHTING PROPORTION Not cycling Cycling or , , , , = 1 adult male present, = >1 adult male present. = 1 adult female present, = >1 adult female present. 59

75 Table II.2. MANOVA results for overall female habitat use selection ratios and the effect of female cycling status and male presence, and overall male habitat use selection ratios and the effect of female presence including cycling status. df = degrees of freedom, MS = mean squares, F = model statistic from the permuted F-tests, R 2 = adjusted R-squared or effect size, P = significance value. df MS F R 2 P Females Cycling Male presence <0.01 Cycling*Male presence Residual Males Female presence Residual

76 CHAPTER III The fitness consequences of early social network strategies in wild bottlenose dolphins 3 INTRODUCTION The key characteristics that define socially complex societies remain controversial (de Waal and Tyack, 2009), but one commonly accepted element is the presence of repeated interactions and individualized relationships (Freeberg et al., 2012). As such, the benefits of social complexity far outweigh simple aggregation or grouping effects, as ephemeral social interactions can develop into lasting relationships or even stable, long-term bonds (Alexander, 1974). The challenges of both group living and complex societies, however, can be great. Individuals compete for access to limited resources, including food, mates, and social position (Silk, 2007a), and are more vulnerable to disease (Parrish and Edelstein-Keshet, 1999). Given these potential costs, how, then, might complex society confer fitness benefits to its constituents? A notable feature of socially complex mammalian societies is slow life histories and a prolonged period of development (Dunbar, 1998; Dunbar, 2009; de Waal and Tyack, 2009). This period is also marked by increased social behavior and is thought to be a period when social skills and social bonds are formed, as well as sex-specific behaviors (Pereira and Fairbanks, 2002). Consequently, the nature and quality of social relationships during the juvenile period might be a good predictor of later success. The relationship between sociality and fitness has been well established in many mammalian systems, from primates to marmots (for a review see 3 Authorship for paper: Megan M. Wallen, Margaret A. Stanton, Ewa Krzyszczyk, & Janet Mann. 61

77 Seyfarth and Cheney, 2012). However, in studies conducted on adults, it is difficult to determine whether sociality confers fitness benefits, or whether high fitness facilitates social relationships. For example, young infants are attractive to other females (e.g., Hrdy, 1976; Mann and Smuts 1998; Silk, 1999) and females with offspring might preferentially associate (Silk et al., 2003a). To date, only few studies have examined how the early social environment relates to fitness (e.g. long-tailed manakins, Chiroxiphia linearis; McDonald, 2007; marmots, Marmota marmota; Berger et al., 2015; geladas, Theropithecus gelada; Barale et al., 2015; baboons, Papio cynocephalus; Tung et al., 2016; horses, Equus caballus; Nuñez et al., 2015; bottlenose dolphins, Tursiops aduncus; Stanton and Mann, 2012; and zebra finches, Taeniopygia guttata; Mariette et al., 2013). These studies suggest that social strategies are evident and critical early on, and help to disentangle the direction of the relationship between sociality and fitness. Same-sex social relationships, such as those between females, have unique benefits (Sterck et al., 1997). In female-philopatric systems, relationships with closely related individuals might increase inclusive fitness benefits (Silk et al., 2006a; b). Social bonds between females may function to mitigate male harassment (Smuts and Smuts; 1993) or reduce stress (Brent et al., 2011), provide learning opportunities such as mothering skills (Lancaster, 1971; Fairbanks, 1993; Mann and Smuts, 1998), and have direct ties to survival and reproductive success (Silk et al., 2003b; 2009; 2010). In the last several decades, network analysis has become an informative tool for understanding the structure and function of complex animal societies. The effects of social interactions extend beyond the dyadic relationships often studied in behavioral ecology, and thus it has become important to characterize the differences in how individuals interact with and influence the larger social network (Brent, 2015). It is well established that variation in social 62

78 network position can have biological significance (see Table III.1 for selected metrics and definitions). For instance, in long-tailed manakins, high information centrality increased the odds that a juvenile male would rise in social rank (McDonald, 2007). Also, Stanton and Mann (2012) found that male bottlenose dolphin calves with high eigenvector centrality had a higher probability of survival as juveniles. Additionally, those males who did not survive had stronger ties to juvenile males than those males who did survive, which may reflect the aggressive and competitive nature of reaching adulthood and forming alliances. Thus, the social strategy that an individual employs in the context of the larger network may have important fitness consequences. Bottlenose dolphins (Tursiops spp.) are long-lived, social mammals that develop strong and stable bonds, exhibit high social heterogeneity, and have high fission-fusion dynamics (Connor et al., 2000). Such fluidity allows for individual variation in group and association preferences. In Shark Bay, adult bottlenose dolphin males form stable bonds between two to three alliance members, who collectively benefit by gaining access to mates (Connor et al, 1992a; b). Additionally, first order alliances may cooperate to form second- and third-order alliances (Connor and Krützen, 2015), increasing the complexity and necessity to form strong male-male bonds. Intriguingly, female sociality is highly variable (Smolker et al., 1992; Mann et al., 2000; Gibson and Mann, 2008a; b). Some females are highly social while others remain more solitary. Female social tendencies may be linked to home range overlap (Frère et al., 2010a), habitat preferences, foraging traditions (Mann et al., 2012), and matrilineal relatedness (Frère et al., 2010a). However, with primarily solitary and kin-based foraging tactics (Mann and Sargeant, 2003; Sargeant et al., 2007; Sargeant and Mann, 2009) with non-defensible resources, and absence of a dominance hierarchy (Connor et al., 2000), there are no clear rules dictating how 63

79 social a female dolphin should be, or which individuals she should associate with. Males clearly benefit from having alliance partners, but it remains less clear why social bonds among females may be beneficial. These qualities make them an ideal system in which to study the fitness effects of individual-level social network position. Through social learning, dolphins learn how to forage (Mann and Sargeant, 2003), raise offspring (Mann and Smuts, 1998), use specific habitats (Sargeant et al., 2007; Mann et al., 2008), and manipulate tools (Mann and Sargeant, 2003; Mann et al., 2008), among other behaviors. Particularly for animals that are philopatric (e.g. Tsai and Mann, 2013), environmental and social conditions remain relatively constant from infancy to maturity, so selection should favor established social bonds that enable relevant information transfer and afford protection from conspecifics and predators. In bottlenose dolphins, social bonds begin to develop during the calf period, prior to weaning (Gibson and Mann, 2008a; b; Stanton et al., 2011), yet little is known about the juvenile period. As mother-calf association drops off post-weaning (Tsai and Mann, 2013), behavior also appears to reflect sex-specific pressures (Daisy Kaplan and Connor, 2007; McHugh et al., 2011; Tsai and Mann, 2013; Krzyszczyk et al., in review). Bottlenose dolphins are not buffered by a stable cohesive kin group as is the case in other long-lived mammals (e.g. primates, Wrangham 1980; killer whales, Bigg, 1990; elephants, Wittemyer et al., 2005), so it is important to develop social strategies from a young age to ease the transition to sexual maturity and integration in the wider social network. Sociality relates to fitness variation in adult female dolphins (Frère et al., 2010b), and in juvenile male dolphins (Stanton and Mann, 2012). In this study, using a social network approach we seek to define and characterize the social strategies that juvenile female dolphins employ and 64

80 relate them to fitness measures. Using 30 years of historical data from the Shark Bay Dolphin Research Project, we explore the adaptive value of individual females early social environment, with the hypothesis that several social strategies may confer fitness benefits. This study is the first to investigate the relationship between early sociality and adult fitness in a cetacean population. By selecting relevant node-level social network metrics (Table III.1), we predict several possible (non-mutually exclusive) strategies to emerge. In the first strategy, reproductively successful females would have high strength, with fairly low degree, emphasizing quality over quantity of social relationships (the best friend strategy). In the second strategy, successful females would have high clustering coefficient, representing a tight social unit where direct associates are also close with each other (the clique strategy). The final strategy we predict is that successful females have high betweenness and/or eigenvector centrality, which suggests well connectedness in the network (the socialite strategy) and the importance of indirect ties. We predict that any one (or all) of these strategies may confer fitness benefits, because females likely adopt individualized strategies. METHODS Study site Shark Bay, Western Australia (25 47 S, E) is home to a resident population of bottlenose dolphins (Tursiops aduncus) that has been studied continuously since

81 Researchers have collected behavioral, ecological, genetic, and demographic data on > 1600 individuals within the 300 km 2 study area on the eastern gulf near Monkey Mia. In particular, mothers and calves have been intensively tracked, enabling detailed study of kin association over time and a more precise estimation of adult female fitness metrics. All individuals are uniquely identified by the shape and damage to the dorsal fin using photo ID, as well as other obvious bodily markings such as tooth rakes and shark bites (Whitehead et al., 2000). Data collection life history and reproduction The juvenile period in females was considered to be from weaning (no longer nursing) until age 11, at which point calculation of calving success began since the earliest known age of pregnancy is 10 years old (J.M., unpublished data). If the date of weaning was unknown, then the juvenile period started at age 4, the average weaning age (Mann et al., 2000). Calving success was calculated as an annual rate of the number of calves that survived to age 3 (because calves that survive until age 3 nearly always survive to weaning; Mann et al., 2000). For example, if a female had 4 surviving calves over 20 reproductive years, her calving success would be 0.2. Only females with 8 consecutive years of known reproductive outcomes were included in this analysis because that time period allows for a female to have at least two surviving calves. Data collection survey association Dolphin association data were calculated from 11,066 surveys conducted from 1985 to Surveys after 2007 are not included because each female had to be surveyed for 8 years as 66

82 an adult for reproductive data. Surveys are brief, five-minute snapshots of group behavior and composition collected by scan sampling (Altmann, 1974; Mann, 1999). Dolphins were considered to be associating if they were within 10 meters of any dolphin in the group in the first five minutes of each survey (Smolker et al., 1992). For this study we focused on the femalefemale networks during the juvenile period, thus networks included females of any age and their dependent calves. For each female with at least 8 years of reproductive outcomes, we selected all surveys from her weaning date to age 11 and used them to create her juvenile social network. Females and their associates were included if they were observed on at least 15 unique days during the juvenile period (following Stanton and Mann, 2012), and only one survey was selected per female per day to reduce temporal and locational dependence. Juvenile females who met both the reproductive outcome threshold and the sighting record threshold yielded a sample of 33 individuals. Data analysis social network metrics All ties between individual dolphins in each social network were weighted by the dyad s half-weight association index (HWI, Cairns and Schwager, 1987). The HWI is given by the equation X/(X + 0.5(Y A + Y B ) + Y AB ), whereby X is the number of days individuals A and B have been seen together, Y A is the number of days A was seen without B, Y B is the number of days B was seen without A, and Y AB is the number of days A and B were both seen, but not together. The HWI is an appropriate measure for the Shark Bay dolphin fission-fusion society because it is less biased when individuals are more likely to be found apart than together (Whitehead, 2008). We 67

83 calculated 5 metrics from each network (Table III.1). All metrics were normalized or scaled to enable comparison across networks of different sizes. Data analysis social strategies and fitness First we ran 5 separate linear models with the social network metric as a fixed effect, and surviving calves weighted by years as the response variable. We did not include all metrics in a single model because they exhibited extreme collinearity (Appendix C, Table S1). To reduce these correlated metrics to independent, aggregated social variables, we used a principal component analysis (PCA). We plotted each juvenile female in the principal component space (using the covariance matrix and a Varimax rotation) to visually distinguish between individual social strategies. The PCA was robust to removal of outliers so all individuals were retained in the final analysis. We then used PC1 and PC2 as fixed predictors in a linear model, with surviving calves weighted by years as the response variable. Finally, we compared the R-squared values of each model (df = 32 for all models) to determine if the PCs were better predictors of calving success than any metric alone. All statistical analyses were conducted in R v3.3.1 (R Core Team, 2016). RESULTS Average calving success (surviving calves per year) across all females was 0.14 (SD = 0.07), and ranged from 0 to Put another way, females successfully calved on average every 68

84 7.14 years. Average HWI of the top associates (97.5 percentile) across all females was 0.19 (SD = 0.09), and values ranged from 0.06 to Across all networks and associates, HWI ranged from 0 (never seen together) to Degree, strength, and clustering coefficient were each significant predictors of calving success when run in separate models, despite high variance (Table III.2, Figures III.1-3). These three metrics also weighed heavily on the first principal component, though eigenvector centrality was weighed heavier than clustering coefficient (Table III.3). Betweenness and clustering coefficient weighed heavily on the second principal component. The first and second principal components explained 55% and 30% of the observed variance, respectively (Table III.4). Though the first three principal components cumulatively explained 94% of the total variance, only the first two principal components were retained for visualization purposes (Figure III.4). In a linear regression, PC1 was a significant predictor of calving success (Table III.5). Because each model had equivalent degrees of freedom, we could directly compare the R- squared values of each model to determine which was best at predicting calving success. The aggregated social strategy represented by PC1 was better at predicting calving success than degree, strength, or clustering coefficient alone (Table III.5, Figure III.5). DISCUSSION This is the first study to demonstrate how early female-female social networks predict female calving success in a cetacean. We examined the juvenile period because of its importance in shaping later life success, particularly in long-lived species with extended development prior 69

85 to sexual maturity. Our results show that three social network metrics, degree, strength and clustering coefficient in the juvenile period, were predictive of calving success in adulthood. While degree, strength, and clustering coefficient all independently predicted calving success, the high variance in the relationships suggests the presence of numerous social tactics; as predicted, there does not appear to be a one-size-fits-all juvenile social trajectory that predicts calving success. Importantly, the first principal component explained the most variation in calving success, suggesting that some combination of degree, strength, and eigenvector centrality during the juvenile period is best at predicting adult fitness. This is a first step in understanding the various social strategies used by female dolphins, and explaining how divergent strategies may be equally beneficial for fitness. The principal components did not clearly differentiate between types of social tactics as expected. The individuals with the highest strength and clustering had the highest fitness, which supports the hypothesis that the strong clique tactic is successful, but other tactics were also successful. Though high success females tended to have higher-than-average centrality, some of the most successful females had the lowest centrality, further supporting the conclusion that connectedness may not be as important as strong dyadic bonds in the female network. In many species, including humans, juveniles or adolescents exhibit age-specific behaviors that often diminish into adulthood. During the period of individual and social development prior to sexual maturity, juveniles often exhibit more exploratory or risk-taking (Spear, 2000) and playful (Fagen, 1977; Barale et al., 2015) behaviors as they learn to independently navigate the social and physical environments. Even though female calves are known to retain patterns of their mother s sociality during separations (Gibson and Mann, 2008a; b) and post-weaning (Mann et al., 2012; Krzyszczyk et al., in review), the high variation in 70

86 juvenile sociality may simply represent of a period of social experimentation before established strategies are developed. As in other species (e.g. forked fungus beetles, Bolitotherus cornutus; Formica et al., 2011), an individual dolphin s position within the network can be viewed as that individual s social phenotype, representing an individualized strategy with a unique combination of social network characteristics. Yet it is still a conundrum why female bottlenose dolphins don t form bonds as a rule, as in other socially complex, long-lived mammals such as baboons (Silk et al., 2003b; Silk et al., 2006; Silk 2007b; Silk et al., 2009; Silk et al., 2010). For example, sponger females have lower degree, strength, and eigenvector centrality, and are generally more solitary, than non-sponger females (Mann et al., 2012), but both groups have similar calving success (Mann et al., 2008). In dolphins, most foraging tactics are solitary (Mann and Sargeant, 2003; Sargeant et al., 2007; Sargeant and Mann, 2009) with non-defensible resources; thus, it may be more beneficial for females to focus their efforts on developing foraging techniques in a heterogeneous environment (Sargeant et al., 2007; Patterson and Mann, 2011), and to reduce competition (Gowans et al., 2007) rather than to develop complex bonds and remain in stable groups. By contrast, adult males in Shark Bay have extremely tight bonds (Mann et al., 2012) and alliance membership is stable across years (Connor and Krützen, 2015). Lone males are unable to monopolize single females; therefore, in order to gain access to reproductive opportunities males have a greater need to develop alliance bonds and cooperate to form consortships with females. Nevertheless, our results suggest that early female social relationships provide some net benefit to juvenile dolphins, and we propose two non-exclusive hypotheses to explain the benefit of female dolphins forming social bonds at a young age. The first is the social support 71

87 hypothesis, which posits that young females form strong bonds with other females to aid with later life demands (Pereira, 1988), such as calf-rearing or calf-socialization opportunities (Gibson and Mann, 2008a; b; Murray et al., 2014), or a counterstrategy to mitigate the impact of aggressive alliances and the costs of association with adult males (Scott et al., 2005; Wallen et al., 2016). Females often engage in affiliative contact behaviors (Connor et al., 2006), which may function to reduce stress during intersexual conflict (Silk, 2002). Female relationships may be a counterstrategy to mitigate the stress and physiological impacts of prolonged herding (e.g. Connor et al., 2006). The second is the learning-to-mother hypothesis, which posits that young females develop strong bonds with lactating females and their dependent calves to develop skills for future successful reproduction (Lancaster, 1971; Fairbanks, 1990; Mann and Smuts, 1998). If juvenile females form such bonds with close kin, they may additionally gain inclusive fitness benefits. It should be noted that these are not mutually exclusive hypotheses; however, biased association with only lactating females would provide stronger support for the learning-tomother hypothesis. Given the relative importance of dyadic bonds versus embeddedness in the wider network, future work characterizing the nature of juvenile female bonds will elucidate the specific function of forming early bonds. Importantly, it is still unknown whether juvenile female dolphins maintain their social characteristics into adulthood (but see geladas, Barale et al., 2015), or whether the juvenile period only represents the early development of individual social tactics. Further, development of strong dyadic social bonds may be equally as important as the development of social skills. Future studies that examine the stability of individual network traits through to adulthood will 72

88 help to disentangle the relative importance of long-term dyadic bonds versus stable social strategies. Several caveats to our study deserve attention. First, the network of each juvenile female was taken over the entire juvenile period (approximately 7 years), when in fact a dynamic network is more realistic. Similarly, we were unable to capture demographic changes and corresponding network adjustments when taking this top-down approach to the juvenile social network. Finally, using survey association for this study was beneficial because it yielded a high number of samples on many individuals. However, we caution against the gambit of the group bias, because individuals in a survey may not have been directly interacting. There are many reasons why two individuals may be sighted together in a survey by chance, such as when large groups of dolphins come together to rest (Heithaus and Dill, 2002) or to forage in large schools of fish (Mann and Sargeant, 2003). However, the use of weighted networks helps to reduce this bias. Reproductive benefits of early sociality imply a delayed fitness payoff. Our study provides evidence for the development of early social strategies that influence adult reproductive success, while reducing the confounding effect that presence of a dependent calf imposes on adult female sociality. A previous study (Frère et al., 2010b) found that adult female social variance was a better predictor of calving success than genetic variance, in that high success females preferred to associate with each other. However, successful females are more likely to be lactating at any given time, and lactating females tend to form larger groups with other lactating females and their calves (Mann and Smuts, 1999; Mann et al., 2000) so the argument is somewhat circular. Our results support Frere et al. in that sociality is an independent predictor of calving success, and that pattern appears to be maintained from the juvenile to the adult period. 73

89 The nature, and thus the specific benefit, of sociality necessarily changes from juvenility to adulthood. For example, juvenile social bonds may be centered on play, while bonds during adulthood may function to protect vulnerable dependent calves. Future work to characterize the quality and composition of the specific relationships developed during juvenility e.g. age-sex homophily, or kin preference will help to elucidate the function of early social bonds. Our results highlight the importance of considering group-wide variation in social network metrics and strategies. In a recent study of rock hyrax (Barocas et al., 2011), the variance (SD) of strength within a social group was a predictor of longevity, however individual strength was not. Therefore, an animal s associates and the similarity of their social preferences may have greater influence throughout the network. Social bonds and social skills are known to be important for fitness in many animals. However, given the difficulty of acquiring longitudinal data on known individuals in wild populations, only few studies have been able to study the relationship between sociality early in life on adult fitness outcomes. Studies that only consider adult traits are missing a key life history stage which is important for understanding behavior and fitness consequences later in life. Further, it is clear that in addition to dyadic-level analysis, network connections and the intragroup variance in such connections, are important to consider cohesively when evaluating the fitness benefits of the social environment at the individual level. 74

90 FIGURES AND TABLES Figure III.1. Effect of normalized degree on calving success. Gray shading indicates 95% confidence interval. Each point represents an individual dolphin s (N = 33) calving success based on their degree as predicted by the linear regression model. 75

91 Figure III.2. Effect of normalized strength on calving success. Gray shading indicates 95% confidence interval. Each point represents an individual dolphin s (N = 33) calving success based on their strength as predicted by the linear regression model. 76

92 Figure III.3. Effect of clustering coefficient on calving success. Gray shading indicates 95% confidence interval. Each point represents an individual dolphin s (N = 33) calving success based on their clustering coefficient as predicted by the linear regression model. 77

93 Figure III.4. Biplot displaying each female (N = 33) in the principal component space and the loadings of each metric on each principal component. 78

94 Figure III.5. Effect of PC1 on calving success. Gray shading indicates 95% confidence interval. Each point represents an individual dolphin s (N = 33) calving success based on the value of their PC1 as predicted by the linear regression model. 79

95 Table III.1. Description of each social network metric analysed in this study, with a comprehensive compilation of studies that have found the metrics to predict direct measures of fitness (including survival and reproduction) or other variables related to fitness. Other variables include social rise (Ryder et al., 2008; Gilby et al., 2013), social status (Glowacki et al., 2016), mating or copulation success (Oh and Badyaev, 2010; Formica et al., 2011), food patch discovery (Aplin et al., 2012), bovine tuberculosis infection (Weber et al., 2013), and serotonin production pathway (Brent et al., 2013). Non-significant results (NS) are also included. Note that other studies may have used association metrics other than the HWI, including the simple ratio index (Wey et al., 2013). Also note that this is not an exhaustive list of all metrics studied with respect to fitness. Other commonly used metrics include information centrality (e.g. McDonald, 2007), closeness (e.g. Weber et al., 2013), and reach (e.g. McDonald, 2007; Cheney et al., 2016). Finally, note that the type of network used to calculate these metrics may vary, including general association networks (this study), affiliative or agonistic interaction networks, or hypothetical preferential gift-giving (Glowaki et al., 2016). 80

96 Metric Description Fitness measure Degree The number of associates for each node, undirected. Independent of the number of interactions between each node. Survival Reproduction Other NS: Bottlenose dolphin g Wire-tailed manakin a ; marmot b NS: degu c ; chimpanzee d Wire-tailed manakin f ; human n ; badger o NS: Long-tailed manakin e ; chimpanzee d Strength Weighted degree, undirected. Takes into account the number of interactions between each node. Bottlenose dolphin g ; Marmot b ; rhesus marmot b macaque h NS: degu c ; chimpanzee d Forked fungus beetle i ; rhesus macaque h NS: chimpanzee d Betweenness centrality A measure of how well connected a node is in the wider social network. Proportion of the shortest paths between any two nodes that pass through that node. NS: Bottlenose dolphin g Chimpanzee d ; wire-tailed manakin a NS: chacma baboon l Long-tailed manakin e ; house finch j ; rhesus macaque h ; tit m (Paridae); chimpanzee d ; badger o Eigenvector centrality A composite measure of how well connected a node and that node s associates are in the global social network. Bottlenose dolphin g ; Wire-tailed manakin a ; bighorn sheep k rhesus macaque h ; bighorn sheep k ; chacma baboon l Long-tailed manakin e ; wire-tailed manakin f ; tit m (Paridae); human n ; rhesus macaque h NS: chimpanzee d NS: chimpanzee d Clustering coefficient A measure of social unit cohesiveness, in which a node s associates are also tight with each other. NS: Bottlenose dolphin g Chacma baboon l Forked fungus beetle i a Ryder et al., b Wey and Blumstein, c Wey et al., d Gilby et al., e McDonald, f Ryder et al., g Stanton and Mann, h Brent et al., i Formica et al., j Oh & Badyaev, k Vander Wal et al., l Cheney et al., m Aplin et al., n Glowacki et al., o Weber et al.,

97 Table III.2. Parameter estimates for fixed effects for the individual linear regression models. Model Estimate ± SE t R 2 P Degree Intercept Normalized degree ± ± * Strength Intercept Normalized strength ± ± < ** Betweenness Eigenvector Clustering coefficient Intercept Normalized betweenness Intercept Scaled eigenvector Intercept Clustering coefficient ± ± ± ± ± ± < < ** Table III.3. Coefficients of the linear combinations of each continuous variable for the first three principal components. PC1 PC2 PC3 Degree Strength Betweenness Eigenvector Clustering coefficient

98 Table III.4. Relative importance of the first three principal components and proportion of variance explained by each principal component. PC1 PC2 PC3 Standard deviation Proportion of variance Cumulative proportion Table III.5. Parameter estimates for fixed effects for the linear regression model on the principal components. The reported R 2 is from the reduced model after removing PC2. Intercept PC1 PC2 Estimate ± SE t R 2 P ± < ± (reduced ** ± model)

99 CONCLUSION Given the fission-fusion social structure, bisexual philopatry, and coercive mating, Shark Bay bottlenose dolphins are an excellent system in which to study the behavioral, ecological, and fitness responses to sexual conflict and diversified social relationships. The costs and benefits of sociality are well known, but in species with long life histories, social complexity, ecological specialization, and extensive social learning, the fitness consequences are context dependent and may be subtle, complex, and/or delayed. During fertile periods, females must balance the tradeoffs between attracting preferred mates, but avoiding suboptimal mates including juvenile males and adult sons. Coercive mating may be a male counterstrategy to restrict female choosiness or mating resistance. As a result of the coercive mating system, female dolphins appear to suffer an ecological cost. Females experience constrained geographic ranging and reduced preferred habitat use when together with males, which may have downstream effects on foraging efficiency and resource acquisition. These constraints are amplified in fertile females. Finally, social skills or social bonds may be one way that females mitigate the effects of coercion. Females appear to develop social relationships from a young age, and the adult fitness benefits of individual social network metrics are already detectable during the juvenile period. Long-term studies such as the Shark Bay Dolphin Research Project are crucial for amassing detailed, individual-level data throughout the lifespan, and can inform fitness measures on both the individual and population level. Increasingly, behavioral studies on wild populations are being used to inform conservation studies and management practices. It is becoming ever more important to conduct 84

100 sustainable, non-invasive research on wild animals to understand how they interact with and influence the surrounding environment. With growing human impact on natural environments, an understanding of species-specific intrinsic stressors will contribute to a fuller picture on the ways in which populations may be vulnerable. This work is the first step to understanding the complex contributions to female fitness with respect to the social environment. 85

101 REFERENCES INTRODUCTION Alexander, R. D. (1974). The evolution of social behavior. Annual Review of Ecology and Systematics, 5, Andersson, M.B. (1994). Sexual selection. Princeton University Press. Arnqvist, G., & Rowe, L. (2013). Sexual conflict. Princeton University Press. Bateman, A. J. (1948). lntra-sexual selection in Drosophila. Heredity, 2(3), Bateson, P. P. G. (1983). Mate choice. Cambridge University Press. Cheney, D. L., Silk, J. B., & Seyfarth, R. M. (2016). Network connections, dyadic bonds and fitness in wild female baboons. Royal Society Open Science, 3(7), Clutton-Brock, T. H., & Parker, G. A. (1995). Sexual coercion in animal societies. Animal Behaviour, 49(5), Clutton-Brock, T.H., & Lukas, D. (2012). The evolution of social philopatry and dispersal in female mammals. Molecular Ecology, 21(3), Connor, R. C., & Krützen, M. (2015). Male dolphin alliances in Shark Bay: changing perspectives in a 30-year study. Animal Behaviour, 103(SI), Freeberg, T. M., Dunbar, R. I., & Ord, T. J. (2012). Social complexity as a proximate and ultimate factor in communicative complexity. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 367(1597),

102 Frère, C., Krützen M., Kopps, A., Ward, P., Mann, J., and Sherwin, W.B. (2010). Inbreeding tolerance and fitness costs in wild bottlenose dolphins. Proceedings of the Royal Society B, 277(1694), Gibson, Q. A., & Mann, J. (2008a). Early social development in wild bottlenose dolphins: sex differences, individual variation and maternal influence. Animal Behaviour, 76(2), Gibson, Q. A., & Mann, J. (2008b). The size, composition and function of wild bottlenose dolphin (Tursiops sp.) mother calf groups in Shark Bay, Australia. Animal Behaviour, 76(2), Krützen, M., Barré, L. M., Connor, R. C., Mann, J., & Sherwin, W. B. (2004). O father: where art thou? Paternity assessment in an open fission fusion society of wild bottlenose dolphins (Tursiops sp.) in Shark Bay, Western Australia. Molecular Ecology, 13(7), Mann, J., Connor, R.C., Barre, L.M., & Heithaus, M.R. (2000). Female reproductive success in bottlenose dolphins (Tursiops sp.): life history, habitat, provisioning, and group-size effects. Behavioral Ecology, 11(2), Mann, J., & Sargeant, B. (2003). Like mother, like calf: the ontogeny of foraging traditions in wild Indian Ocean bottlenose dolphins (Tursiops sp.). In The Biology of Traditions (Eds. D. M. Fragaszy and S. Perry), pp Cambridge, MA, U.S.A.: Cambridge University Press. Mann, J., Sargeant, B. L., Watson-Capps, J. J., Gibson, Q. A., Heithaus, M. R., Connor, R. C., & Patterson, E. (2008). Why do dolphins carry sponges? PLoS One, 3(12), e

103 Mann, J., Stanton, M. A., Patterson, E. M., Bienenstock, E. J., & Singh, L. O. (2012). Social networks reveal cultural behaviour in tool-using dolphins. Nature Communications, 3, 980. Mateo, J. M. (2004). Recognition systems and biological organization: the perception component of social recognition. Annales Zoologici Fennici, 41(6), Patterson, E. M., & Mann, J. (2011). The ecological conditions that favor tool use and innovation in wild bottlenose dolphins (Tursiops sp.). PLoS One, 6(7), e Patterson, E. M., Krzyszczyk, E., & Mann, J. (2016). Age-specific foraging performance and reproduction in tool-using wild bottlenose dolphins. Behavioral Ecology, 27(2), Sargeant, B. L., Wirsing, A. J., Heithaus, M. R., & Mann, J. (2007). Can environmental heterogeneity explain individual foraging variation in wild bottlenose dolphins (Tursiops sp.)?. Behavioral Ecology and Sociobiology, 61(5), Sargeant, B. L., & Mann, J. (2009). Developmental evidence for foraging traditions in wild bottlenose dolphins. Animal Behaviour, 78I(3), Scott, E., Mann, J., & Watson-Capps, J. J. (2005). Aggression in bottlenose dolphins: evidence for sexual coercion, male-male competition, and female tolerance through analysis of tooth-rake marks and behaviour. Behaviour, 142(1), Seyfarth, R. M., & Cheney, D. L. (2012). The evolutionary origins of friendship. Annual Review of Psychology, 63, Silk, J. B., Beehner, J. C., Bergman, T. J., Crockford, C., Engh, A. L., Moscovice, L. R., Wittig, R. M., Seyfarth, R. M., & Cheney, D. L. (2009). The benefits of social capital: close social bonds among female baboons enhance offspring survival. Proceedings of the Royal Society B: Biological Sciences, 276(1670),

104 Silk, J. B., Beehner, J. C., Bergman, T. J., Crockford, C., Engh, A. L., Moscovice, L. R., Wittig, R. M., Seyfarth, R. M., & Cheney, D. L. (2010). Strong and consistent social bonds enhance the longevity of female baboons. Current Biology, 20(15), Smolker, R. A., Richards, A. F., Connor, R. C., & Pepper, J. W. (1992). Sex differences in patterns of association among Indian Ocean bottlenose dolphins. Behavior, 123(1), Smuts, B. B., & Smuts, R. W. (1993). Male aggression and sexual coercion of females in nonhuman primates and other mammals: evidence and theoretical implications. Advances in the Study of Behavior, 22(22), Stanton, M.A., Gibson, Q.A., & Mann, J. (2011). When mum s away: a study of mother and calf ego networks during separations in wild bottlenose dolphins (Tursiops sp.). Animal Behaviour, 82(2), Stanton, M.A., & Mann, J. (2012). Early social networks predict survival in wild bottlenose dolphins. PloS One, 7(10), e Tsai, Y.J., & Mann, J. (2013). Dispersal, philopatry, and the role of fission-fusion dynamics in bottlenose dolphins. Marine Mammal Science, 29(2), Wallen, M. M., Patterson, E. M., Krzyszczyk, E., & Mann, J. (2016). The ecological costs to females in a system with allied sexual coercion. Animal Behaviour, 115, Watson-Capps J. J. (2005). Female mating behavior in the context of sexual coercion and female ranging behavior of bottlenose dolphins (Tursiops sp.) in Shark Bay, Western Australia. Ph.D. dissertation, Georgetown University. 89

105 Watson-Capps, J. J. (2009). Evolution of sexual coercion with respect to sexual selection and sexual conflict theory. In Sexual Coercion in Primates and Humans (Eds. M. N. Muller and R. W. Wrangham), pp Harvard University Press, Cambridge, Massachusetts. Wrangham, R. W. (1980). An ecological model of female-bonded primate groups. Behaviour, 75(3), CHAPTER I Alberts, S. C., Watts, H. E., & Altmann, J. (2003). Queuing and queue-jumping: long-term patterns of reproductive skew in male savannah baboons, Papio cynocephalus. Animal Behaviour, 65(4), Altmann, J. (1974). Observational study of behavior: sampling methods. Behaviour, 49(3), Álvarez, G., Ceballos, F. C., & Berra, T. M. (2015). Darwin was right: inbreeding depression on male fertility in the Darwin family. Biological Journal of the Linnean Society, 114(2), Amaral, R. S. (2010). Use of alternative matrices to monitor steroid hormones in aquatic mammals: a review. Aquatic Mammals, 36(2), 162. Amos, B., Schlötterer, C., & Tautz, D. (1993). Social structure of pilot whales revealed by analytical DNA profiling. Science, 260(5108), Aureli, F., Schaffner, C. M., Boesch, C., Bearder, S. K., Call, J., Chapman, C. A., Connor, R., Di Fiore, A., Dunbar, R. I. M., Henzi, S. P., Holekamp, K., Korstjens, A. H., Layton, R., 90

106 Lee, P., Lehmann, J., Manson, J. H., Ramos-Fernandez, G., Strier, K. B., & van Schaik, C. P. (2008). Fission fusion dynamics. Current Anthropology, 49(4), Ayres, K.L., Booth, R.K., Hempelmann, J.A., Koski, K.L., Emmons, C.K., Baird, R.W., Balcomb-Bartok, K., Hanson, M.B., Ford, M.J. and Wasser, S.K., Distinguishing the impacts of inadequate prey and vessel traffic on an endangered killer whale (Orcinus orca) population. PLoS One, 7(6), p.e Baird, R. W., & Whitehead, H. (2000). Social organization of mammal-eating killer whales: group stability and dispersal patterns. Canadian Journal of Zoology, 78(12), Bergfelt, D. R., Steinetz, B. G., Reif, J. S., Schaefer, A. M., Bossart, G. D., Mazzoil, M. S., Zolman, E., & Fair, P.A. (2013). Evaluation of single-sample analysis of progesterone in combination with relaxin for diagnosis of pregnancy in wild bottlenose dolphins (Tursiops truncatus). Aquatic Mammals, 39(2), Biancani, B., Da Dalt, L., Lacave, G., Romagnoli, S., & Gabai, G. (2009). Measuring fecal progestogens as a tool to monitor reproductive activity in captive female bottlenose dolphins (Tursiops truncatus). Theriogenology, 72(9), Brook, F. M., Kinoshita, R., Brown, B., & Metreweli, C. (2000). Ultrasonographic imaging of the testis and epididymis of the bottlenose dolphin, Tursiops truncatus aduncas. Journal of Reproduction and Fertility, 119(2), Brown, J. L., & Eklund, A. (1994). Kin recognition and the major histocompatibility complex: an integrative review. American Naturalist, 143(3), Clutton-Brock, T.H., & Lukas, D. (2012). The evolution of social philopatry and dispersal in female mammals. Molecular Ecology, 21(3),

107 Connor, R. C., Smolker, R. A., and Richards, A. F. (1992a). Two levels of alliance formation among male bottlenose dolphins (Tursiops sp.). Proceedings of the National Academy of Sciences of the United States of America, 89, Connor, R. C., Smolker, R. A., and Richards, A. F. (1992b). Dolphin Alliances and Coalitions. In A. H. Harcourt and F. B. M de Waal (Eds.) Coalitions and Alliances in Humans and Other Animals (pp ), New York, NY, U.S.A.: Oxford University Press. Connor, R.C., Richards, A., Smolker, R., & Mann, J. (1996). Patterns of female attractiveness in Indian Ocean bottlenose dolphins. Behaviour, 133(1), Connor, R. C. and Krützen, M. (2015). Male dolphin alliances in Shark Bay: changing perspectives in a 30-year study. Animal Behaviour, 103, Cooke, S. J., Blumstein, D. T., Buchholz, R., Caro, T., Fernández-Juricic, E., Franklin, C. E.,... & Wikelski, M. (2014). Physiology, behavior, and conservation. Physiological and Biochemical Zoology, 87(1), Cords, M., & Mann, J. (2014). Social conflict management in primates: Is there a case for dolphins? In Primates and Cetaceans, pp Springer Japan. Darwin C The variation of animals and plants under domestication. London: John Murray. Darwin C The effects of cross and self fertilization in the vegetable kingdom. London: John Murray. Deecke, V. B., Barrett-Lennard, L. G., Spong, P., & Ford, J. K. (2010). The structure of stereotyped calls reflects kinship and social affiliation in resident killer whales (Orcinus orca). Naturwissenschaften, 97(5), Dobson, F. S. (1982). Competition for mates and predominant juvenile male dispersal in mammals. Animal Behaviour, 30(4),

108 Dunn, D.G., Barco, S.G., Pabst, D.A., & McLellan, W.A. (2002). Evidence for infanticide in bottlenose dolphins of the western North Atlantic. Journal of Wildlife Diseases, 38(3), Frère, C., Krützen M., Kopps, A., Ward, P., Mann, J., and Sherwin, W.B. (2010a). Inbreeding tolerance and fitness costs in wild bottlenose dolphins. Proceedings of the Royal Society B, 277(1694), Frère, C.H., Krützen, M., Mann, J., Watson-Capps, J.J., Tsai, Y.J., Patterson, E.M., Connor, R.C., Bejder, L., and Sherwin, W.B. (2010b). Home range overlap, matrilineal and biparental kinship drive female associations in bottlenose dolphins. Animal Behaviour, 80(3), Frère, C. H., Krützen, M., Mann, J., Connor, R. C., Bejder, L., & Sherwin, W. B. (2010c). Social and genetic interactions drive fitness variation in a free-living dolphin population. Proceedings of the National Academy of Sciences, 107(46), Gero, S., Bejder, L., Whitehead, H., Mann, J., & Connor, R. C. (2005). Behaviourally specific preferred associations in bottlenose dolphins, Tursiops spp. Canadian Journal of Zoology, 83(12), Gibson, Q. A., & Mann, J. (2008a). Early social development in wild bottlenose dolphins: sex differences, individual variation and maternal influence. Animal Behaviour, 76(2), Gibson, Q. A., & Mann, J. (2008b). The size, composition and function of wild bottlenose dolphin (Tursiops sp.) mother calf groups in Shark Bay, Australia. Animal Behaviour, 76(2),

109 Goymann, W. (2012). On the use of non-invasive hormone research in uncontrolled, natural environments: the problem with sex, diet, metabolic rate and the individual. Methods in Ecology and Evolution, 3(4), Heithaus, M., & Dill, L. (2002). Food availability and tiger shark predation risk influence bottlenose dolphin habitat use. Ecology, 83, Hildebrandt, T. B., Lueders, I., Hermes, R., Goeritz, F., & Saragusty, J. (2011). Reproductive cycle of the elephant. Animal Reproduction Science, 124(3), Kaplan, J., Lentell, B.J., & Lange, W. (2009). Possible evidence for infanticide among bottlenose dolphins (Tursiops truncatus) off St. Augustine, Florida. Marine Mammal Science, 25(4), Karniski, C., Patterson, E. M., Krzyszczyk, E., Foroughirad, V., Stanton, M. A., & Mann, J. (2015). A comparison of survey and focal follow methods for estimating individual activity budgets of cetaceans. Marine Mammal Science, 31(3), Krützen, M., Barré, L. M., Connor, R. C., Mann, J., & Sherwin, W. B. (2004). O father: where art thou? Paternity assessment in an open fission fusion society of wild bottlenose dolphins (Tursiops sp.) in Shark Bay, Western Australia. Molecular Ecology, 13(7), Lacave, G., Eggermont, M., Verslycke, T., Brook, F., Salbany, A., Roque, L. and Kinoshita, R., Prediction from ultrasonographic measurements of the expected delivery date in two species of bottlenosed dolphin (Tursiops truncatus and Tursiops aduncus). The Veterinary Record, 154(8), pp Lu, A., Beehner, J.C., Czekala, N.M., & Borries, C. (2012). Juggling priorities: female mating tactics in Phayre s leaf monkeys. American Journal of Primatology, 74(5),

110 Mann, J. (1999). Behavioral sampling methods for cetaceans: a review and critique. Marine Mammal Science, 15(1), Mann, J. and Smuts, B. (1999). Behavioral development in wild bottlenose dolphin newborns (Tursiops sp). Behaviour, 136(5), Mann, J., Connor, R.C., Barre, L.M., & Heithaus, M.R. (2000). Female reproductive success in bottlenose dolphins (Tursiops sp.): life history, habitat, provisioning, and group-size effects. Behavioral Ecology, 11(2), Mann, J., & Sargeant, B. (2003). Like mother, like calf: the ontogeny of foraging traditions in wild Indian Ocean bottlenose dolphins (Tursiops sp.). In The Biology of Traditions (Eds. D. M. Fragaszy & S. Perry), pp Cambridge, MA. Cambridge University Press. Mann, J. (2006). Establishing trust: socio-sexual behaviour and the development of male-male bonds among Indian Ocean bottlenose dolphins. Homosexual behaviour in animals, Mateo, J. M. (2004). Recognition systems and biological organization: the perception component of social recognition. Annales Zoologici Fennici, 41(6), McHugh, K. A., Allen, J. B., Barleycorn, A. A., & Wells, R. S. (2011). Natal philopatry, ranging behavior, and habitat selection of juvenile bottlenose dolphins in Sarasota Bay, Florida. Journal of Mammalogy, 92(6), Millspaugh, J. J., & Washburn, B. E. (2004). Use of fecal glucocorticoid metabolite measures in conservation biology research: considerations for application and interpretation. General and Comparative Endocrinology, 138(3), Möller, L.M., & Beheregaray, L.B. (2004). Genetic evidence for sex-biased dispersal in resident bottlenose dolphins (Tursiops aduncus). Molecular Ecology, 13(6),

111 Moore, J., & Ali, R. (1984). Are dispersal and inbreeding avoidance related? Animal Behaviour, 32(1), Muggeo, V. M. (2003). Estimating regression models with unknown break-points. Statistics in Medicine, 22(19), Muggeo, V. M. (2008). Segmented: an R package to fit regression models with broken-line relationships. R news, 8(1), Muller, M. N., Thompson, M. E., Kahlenberg, S. M., & Wrangham, R. W. (2011). Sexual coercion by male chimpanzees shows that female choice may be more apparent than real. Behavioral Ecology and Sociobiology, 65(5), Muraco, H.S. (2015). Reproductive biology of the female bottlenose dolphin (Tursiops truncatus). Ph.D. Dissertation, Mississippi State University. Muraco, H., & Kuczaj, S.A. (2015). Conceptive estrus behavior in three bottlenose dolphins (Tursiops truncatus). Animal Behavior and Cognition, 2(1), O'Brien, J. K. and Robeck, T. R. (2012). The relationship of maternal characteristics and circulating progesterone concentrations with reproductive outcome in the bottlenose dolphin (Tursiops truncatus) after artificial insemination, with and without ovulation induction, and natural breeding. Theriogenology, 78, Ottensmeyer, C.A., & Whitehead, H. (2003). Behavioural evidence for social units in longfinned pilot whales. Canadian Journal of Zoology, 81(8), Patterson, I.A.P., Reid, R.J., Wilson, B., Grellier, K., Ross, H.M., & Thompson, P.M. (1998). Evidence for infanticide in bottlenose dolphins: an explanation for violent interactions with harbour porpoises? Proceedings of the Royal Society of London B, 265(1402),

112 Penn, D. J. (2002). The scent of genetic compatibility: sexual selection and the major histocompatibility complex. Ethology, 108(1), Perez, S., Garcia-Lopez, A., De Stephanis, R., Gimenez, J., Garcia-Tiscar, S., Verborgh, P., Mancera, J. M., & Martinez-Rodriguez, G. (2011). Use of blubber levels of progesterone to determine pregnancy in free-ranging live cetaceans. Marine Biology, 158(7), Perrtree, R.M., Sayigh, L.S., Williford, A., Bocconcelli, A., Curran, M.C., & Cox, T.M. (2016). First observed wild birth and acoustic record of a possible infanticide attempt on a common bottlenose dolphin (Tursiops truncatus). Marine Mammal Science, 32(1), Preis, A., Mugisha, L., Hauser, B., Weltring, A., & Deschner, T. (2011). Androgen and androgen metabolite levels in serum and urine of East African chimpanzees (Pan troglodytes schweinfurthii): Comparison of EIA and LC MS analyses. General and Comparative Endocrinology, 174(3), Pusey, A. E. and C. Packer. (1987). Dispersal and philopatry. In: Primate Societies (Eds. B.B. Smuts, D.L. Cheney, R.M. Seyfarth, R.W. Wrangham, T.T. Struhsaker), pp Chicago: University of Chicago Press. Pusey, A., & Wolf, M. (1996). Inbreeding avoidance in animals. Trends in Ecology and Evolution, 11(5), R Core Team. (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL Read, A.J., Wells, R.S., Hohn, A.A. and Scott, M.D., Patterns of growth in wild bottlenose dolphins, Tursiops truncatus. Journal of Zoology, 231(1),

113 Robeck, T.R., Schneyer, A.L., McBain, J.F., Dalton, L.M., Walsh, M.T., Czekala, N.M., & Kraemer, D.C. (1993). Analysis of urinary immunoreactive steroid metabolites and gonadotropins for characterization of the estrous cycle, breeding period, and seasonal estrous activity of captive killer whales (Orcinus orca). Zoo Biology, 12(2), Robeck, T.R., Steinman, K.J., Yoshioka, M., Jensen, E., O Brien, J.K., Katsumata, E., Gili, C., McBain, J.F., Sweeney, J. and Monfort, S.L. (2005). Estrous cycle characterisation and artificial insemination using frozen thawed spermatozoa in the bottlenose dolphin (Tursiops truncatus). Reproduction, 129(5), pp Robinson, K. P. (2014). Agonistic intraspecific behavior in free-ranging bottlenose dolphins: Calf-directed aggression and infanticidal tendencies by adult males. Marine Mammal Science, 30(1), Rolland, R.M., Hunt, K.E., Kraus, S.D. and Wasser, S.K., Assessing reproductive status of right whales (Eubalaena glacialis) using fecal hormone metabolites. General and Comparative Endocrinology, 142(3), pp Sargeant, B. L., Wirsing, A. J., Heithaus, M. R., & Mann, J. (2007). Can environmental heterogeneity explain individual foraging variation in wild bottlenose dolphins (Tursiops sp.)? Behavioral Ecology and Sociobiology, 61(5), Sargeant, B. L. and Mann, J. (2009). Developmental evidence for foraging traditions in wild bottlenose dolphins. Animal Behaviour, 78(3), Schroeder, P.J. (1990). Breeding bottlenose dolphins in captivity. In: The Bottlenose Dolphin (Eds. S. Leatherwood & R.R. Reeves). pp Academic Press, New York. 98

114 Scott, E., Mann, J., & Watson-Capps, J. J. (2005). Aggression in bottlenose dolphins: evidence for sexual coercion, male-male competition, and female tolerance through analysis of tooth-rake marks and behaviour. Behaviour, 142(1), Smolker, R. A., Richards, A. F., Connor, R. C., & Pepper, J. W. (1992). Sex differences in patterns of association among Indian Ocean bottlenose dolphins. Behavior, 123(1), Stanton, M.A., & Mann, J. (2012). Early social networks predict survival in wild bottlenose dolphins. PloS One, 7(10), e Szulkin, M., Stopher, K. V., Pemberton, J. M., & Reid, J. M. (2013). Inbreeding avoidance, tolerance, or preference in animals? Trends in Ecology and Evolution, 28(4), Tsai, Y.J., & Mann, J. (2013). Dispersal, philopatry, and the role of fission-fusion dynamics in bottlenose dolphins. Marine Mammal Science, 29(2), van Schaik, C. P., Pradhan, G. R., & van Noordwijk, M. A. (2004). Mating conflict in primates: infanticide, sexual harassment and female sexuality. In Sexual selection in primates: new and comparative perspectives (Eds. P. M. Kappeler & C. P. Van Schaik), pp Cambridge University Press, Cambridge. Wallen, M. M., Patterson, E. M., Krzyszczyk, E., & Mann, J. (2016). The ecological costs to females in a system with allied sexual coercion. Animal Behaviour, 115, Watson-Capps J. J Female mating behavior in the context of sexual coercion and female ranging behavior of bottlenose dolphins (Tursiops sp.) in Shark Bay, Western Australia. Ph.D. dissertation, Georgetown University. Watts, D. P. (2015). Mating behavior of adolescent male chimpanzees (Pan troglodytes) at Ngogo, Kibale National Park, Uganda. Primates, 56(2),

115 Wells, R. S., Scott, M. D., & Irvine, A. B. (1987). The social structure of free-ranging bottlenose dolphins. In Current Mammalogy (pp ). Springer US. Wells, R. S., & Scott, M. D. (1990). Estimating bottlenose dolphin population parameters from individual identification and capture-release techniques. In: Individual Recognition of Cetaceans: Use of Photo-Identification and Other Techniques to Estimate Population Parameters. Report of the International Whaling Commission, Special Issue 12, (Eds. P.S. Hammond, S. A. Mizroch, and G. P Donovan), pp Cambridge, UK. Wells, R., (1991). The role of long-term study in understanding the social structure of a bottlenose dolphin community. In Dolphin Societies: Discoveries and Puzzles (Eds. K Pryor & K.S. Norris), pp Berkeley: University of California Press. Wells, R. S., Rhinehart, H. L., Hansen, L. J., Sweeney, J. C., Townsend, F. I., Stone, R., Casper, D. R., Scott, M. D., Hohn, A. A., & Rowles, T. K. (2004). Bottlenose dolphins as marine ecosystem sentinels: developing a health monitoring system. EcoHealth, 1(3), Wright, B. M., Stredulinsky, E. H., Ellis, G. M., & Ford, J. K. B. (2016). Kin-directed food sharing promotes lifetime natal philopatry of both sexes in a population of 2 fish-eating killer whales (Orcinus orca). Animal Behaviour, 115, Whitehead, H., Christal, J., & Tyack, P.L. (2000). Studying cetacean social structure in space and time. In: Cetacean Societies (Eds. J. Mann, R.C. Connor, P.L. Tyack, & H. Whitehead), pp The University of Chicago Press: Chicago. Xu, R. (2003). Measuring explained variation in linear mixed effects models. Statistics in Medicine, 22(22),

116 Yuen, W.H.Q., An assessment of reproductive development of the male Indo-Pacific bottlenose dolphin, Tursiops aduncus, in captivity. Ph.D. dissertation, The Hong Kong Polytechnic University. CHAPTER II Aloise King, E. D., Banks, P. B., and Brooks, R. C. (2013). Sexual conflict in mammals: consequences for mating systems and life history. Mammal Review, 43, Altmann, J. (1974). Observational study of behavior: sampling methods. Behaviour, 49, Anderson, M. J. (2001). Permutation tests for univariate or multivariate analysis of variance and regression. Canadian Journal of Fisheries and Aquatic Sciences, 58, Bates, D., Maechler, M., Bolker, B., and Walker, S. (2014). lme4: Linear mixed-effects models using Eigen and S4. R package version Cartwright, R. and Sullivan, M. (2009). Associations with multiple male groups increase the energy expenditure of humpback whale (Megaptera novaeangliae) female and calf pairs on the breeding grounds. Behaviour, 146, del Castillo, S. M., Bashaw, M. J., Patton, M. L., Rieches, R. R., and Bercovitch, F. B. (2005). Fecal steroid analysis of female giraffe (Giraffa camelopardalis) reproductive condition and the impact of endocrine status on daily time budgets. General and Comparative Endocrinology, 141, Cheal, A.J. and Gales, N.J. (1992). Growth, sexual maturity and food-intake of Australian Indian Ocean bottlenose dolphins, Tursiops-truncatus, in captivity. Australian Journal of Zoology, 40,

117 Clutton-Brock, T. H. and Parker, G. A. (1995). Sexual coercion in animal societies. Animal Behaviour, 49, Connor, R. C., Smolker, R. A., and Richards, A. F. (1992a). Two levels of alliance formation among male bottlenose dolphins (Tursiops sp.). Proceedings of the National Academy of Sciences of the United States of America, 89, Connor, R. C., Smolker, R. A., and Richards, A. F. (1992b). Dolphin Alliances and Coalitions. In A. H. Harcourt and F. B. M de Waal (Eds.) Coalitions and Alliances in Humans and Other Animals (pp ), New York, NY, U.S.A.: Oxford University Press. Connor, R. C., Richards, A., Smolker, R., and Mann, J. (1996). Patterns of female attractiveness in Indian Ocean bottlenose dolphins. Behaviour, 133, Connor, R. C. and Smolker, R. (1996). Pop goes the dolphin: a vocalization male bottlenose dolphins produce during consortships. Behaviour, 133, Connor, R. C., Heithaus, M. R., Barre, L. M., Spoor, F., Higgins, P. O., Dean, C., and Lieberman, D. E. (1999). Superalliance of bottlenose dolphins. Nature, 397, Connor, R. C., Heithaus, M. R., and Barre, L. M. (2001). Complex social structure, alliance stability and mating access in a bottlenose dolphin super-alliance. Proc. R. Soc. Lond. B, 268, Connor, R. C. and Whitehead, H. (2005). Alliances II: rates of encounter during resource utilization: a general model of intrasexual alliance formation in fission-fusion societies. Animal Behaviour, 69, Connor, R. C., Mann, J., and Watson-Capps, J. J. (2006). A sex-specific affiliative contact behavior in Indian Ocean bottlenose dolphins, Tursiops sp. Ethology, 112,

118 Connor, R. C. and Vollmer, N. L. (2009). Sexual coercion in dolphin consortships: a comparison with chimpanzees. In M. N. Muller and R. W. Wrangham (Eds.), Sexual Coercion in Primates and Humans (pp ), Cambridge, MA, U.S.A.: Harvard University Press. Connor, R. C., Watson-Capps, J. J., Sherwin, W. S., and Krützen, M. (2011). A new level of complexity in the male alliance networks of Indian Ocean bottlenose dolphins (Tursiops sp.). Biology Letters, 7, Connor, R. C. and Krützen, M. (2015). Male dolphin alliances in Shark Bay: changing perspectives in a 30-year study. Animal Behaviour, 103, Darden, S. K. and Croft, D. P. (2008). Male harassment drives females to alter habitat use and leads to segregation of the sexes. Biology Letters, 4, Darden, S. K., James, R., Ramnarine, I. W., and Croft, D. P. (2009). Social implications of the battle of the sexes: sexual harassment disrupts female sociality and social recognition. Proceedings of the Royal Society B, 276, Darden, S. K. and Watts, L. (2012). Male sexual harassment alters female social behaviour towards other females. Biology Letters, 8, Feldblum, J. T., Wroblewski, E. E., Rudicell, R. S., Hahn, B. H., Paiva, T., Cetinkaya-Rundel, M., Pusey, A., and Gilby, I. C. (2014). Sexually coercive male chimpanzees sire more offspring. Current Biology, 24, Frère, C., Krützen M., Kopps, A., Ward, P., Mann, J., and Sherwin, W. B. (2010a). Inbreeding tolerance and fitness costs in wild bottlenose dolphins. Proceedings of the Royal Society B, 277, Frère, C. H., Krützen, M., Mann, J., Watson-Capps, J. J., Tsai, Y. J., Patterson, E. M., Connor, R.C., Bejder, L., and Sherwin, W. B. (2010b). Home range overlap, matrilineal and 103

119 biparental kinship drive female associations in bottlenose dolphins. Animal Behaviour, 80, Furuichi, T., Connor, R. C., and Hashimoto, C. (2014). Non-conceptive sexual interactions in monkeys, apes & toothed whales. In J. Yamagiwa and L. Karczmarski (Eds.) Primates & cetaceans: Field research and conservation of complex mammalian societies (pp ). Tokyo, Japan: Springer. Galimberti, F., Boitani, L., and Marzetti, I., (2000). The frequency and costs of harassment in southern elephant seals. Ethology Ecology & Evolution, 12, Gay, L., Eady, P. E., Vasudev, R., Hosken, D. J., and Tregenza, T. (2009). Costly sexual harassment in a beetle. Physiological Entomology, 34, Gibson, Q. A. and Mann, J. (2008). The size, composition and function of wild bottlenose dolphin (Tursiops sp.) mother-calf groups in Shark Bay, Australia. Animal Behaviour, 76, Hasegawa, T. (1990). Sex differences in ranging patterns. In T. Nishida (Ed.) The Chimpanzees of the Mahale Mountains (pp ), Japan: University of Tokyo Press. Heithaus, M. and Dill, L. (2002). Food availability and tiger shark predation risk influence bottlenose dolphin habitat use. Ecology, 83, Heithaus, M.R. and Dill, L.M. (2006). Does tiger shark predation risk influence foraging habitat use by bottlenose dolphins at multiple spatial scales? Oikos, 114, Heubel, K. U. and Plath, M. (2008). Influence of male harassment and female competition on female feeding behaviour in a sexual-asexual mating complex of mollies (Poecilia mexicana, P. formosa). Behavioral Ecology and Sociobiology, 62,

120 Hiruki, L. M., Stirling, I., Gilmartin, W. G., Johanos, T. C., and Becker, B. L. (1993). Significance of wounding to female reproductive success in Hawaiian monk seals (Monachus schauinslandi) at Laysan Island. Canadian Journal of Zoology, 71, den Hollander, M. and Gwynne, D. T. (2009). Female fitness consequences of male harassment and copulation in seed beetles, Callosobruchus maculatus. Animal Behaviour, 78, Hothorn, T., Hornik, K., van de Wiel, M. A., and Zeileis, A. (2006). A Lego system for conditional inference. The American Statistician, 60, Karniski, C., Patterson, E. M., Krzyszczyk, E., Foroughirad, V., Stanton, M., and Mann, J. (2015). A comparison of survey and focal follow methods for estimating individual activity budgets of cetaceans. Marine Mammal Science. DOI: /mms Köhler, A., Hildenbrand, P., Schleucher, E., Riesch, R., Arias-Rodriguez, L., Streit, B., and Plath, M. (2011). Effects of male sexual harassment on female time budgets, feeding behavior, and metabolic rates in a tropical livebearing fish (Poecilia mexicana). Behavioral Ecology and Sociobiology, 65, Krützen, M., Sherwin, W. B., Berggren, P., and Gales, N. (2004). Population structure in an inshore cetacean revealed by microsatellite and mtdna analysis: bottlenose dolphins (Tursiops sp.) in Shark Bay, Western Australia. Marine Mammal Science, 20, Krzyszczyk, E. and Mann, J. (2012). Why become speckled? Ontogeny and function of speckling in Shark Bay bottlenose dolphins (Tursiops sp.). Marine Mammal Science, 28,

121 Link, A., De Fiore, A., and Spehar, S. N. (2009). Female-directed aggression and social control in spider monkeys. In M. N. Muller and R. W. Wrangham (Eds.), Sexual Coercion in Primates and Humans (pp ), Cambridge, MA, U.S.A.: Harvard University Press. Le Boeuf, B. and Mesnick, S. (1991). Sexual behavior of male northern elephant seals: I. Lethal injuries to adult females. Behaviour, 116, Manly, B. F. J., Mcdonald, L. L., Thomas, D. L., McDonald, T. L. and Erickson, W. P Resource Selection by Animals Statistical Design and Analysis for Field Second Edition. 2nd edn. Dordrecht: Kluwer Academic Press. Mann, J. (1999). Behavioral sampling methods for cetaceans: a review and critique. Marine Mammal Science, 15, Mann, J. and Smuts, B. (1999). Behavioral development in wild bottlenose dolphin newborns (Tursiops sp). Behaviour, 136, Mann, J., Connor, R. C., Barre, L. M., and Heithaus, M. R. (2000). Female reproductive success in bottlenose dolphins (Tursiops sp.): life history, habitat, provisioning, and group-size effects. Behavioral Ecology, 11, Mann, J., and Sargeant, B. (2003). Like mother, like calf: the ontogeny of foraging traditions in wild Indian ocean bottlenose dolphins (Tursiops sp.). In D. M. Fragaszy and S. Perry (Eds.), The Biology of Traditions (pp ), Cambridge, MA, U.S.A.: Cambridge University Press. Mann, J., Sargeant, B. L., Watson-Capps, J. J., Gibson, Q. A., Heithaus, M. R., Connor, R. C., and Patterson, E. (2008). Why do dolphins carry sponges? PloS ONE, 3, e

122 Mann, J., Stanton, M. A., Patterson, E. M., Bienenstock, E. J., and Singh, L. O. (2012). Social networks reveal cultural behaviour in tool-using dolphins. Nature Communications, 3, 980. Mann, J. and Patterson, E. M. (2013). Tool use by aquatic animals. Philosophical Transactions of the Royal Society B, 368, Matson, T.K., Putland, D. A., Jarman, P. J., le Roux, J., and Goldizen, A. W. (2007). Influences of parturition on home range and microhabitat use of female black-faced impalas. Journal of Zoology, 271, Matsumoto-Oda, A., and Oda, R. (1998). Changes in the activity budget of cycling female chimpanzees. American Journal of Primatology, 46, Muller, M. N., Kahlenberg, S. M., Thompson, M. E., and Wrangham, R. W. (2007). Male coercion and the costs of promiscuous mating for female chimpanzees. Proceedings of the Royal Society B, 274, Muller, M. N., Kahlenberg, S. M., and Wrangham, R. W. (2009). Male aggression against females and sexual coercion in chimpanzees. In M. N. Muller and R. W. Wrangham (Eds.), Sexual Coercion in Primates and Humans (pp ), Cambridge, MA, U.S.A.: Harvard University Press. Noë, R. (1992). Alliance formation among male baboons: shopping for profitable partners. In A. H. Harcourt and F. B. M de Waal (Eds.) Coalitions and Alliances in Humans and Other Animals (pp ), New York, NY, U.S.A.: Oxford University Press. O'Brien, J. K. and Robeck, T. R. (2012). The relationship of maternal characteristics and circulating progesterone concentrations with reproductive outcome in the bottlenose 107

123 dolphin (Tursiops truncatus) after artificial insemination, with and without ovulation induction, and natural breeding. Theriogenology, 78, Ojanguren, A. F., and Magurran, A. E. (2007). Male harassment reduces short-term female fitness in guppies. Behaviour, 144, Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., O'Hara, R.B., Simpson, G. L., Solymos, P., Stevens, M. H. H. and Wagner, H. (2013). vegan: Community ecology package. R package version Owen, E., Wells, R. S. and Hofmann, S. (2002). Ranging and association patterns of paired and unpaired adult male Atlantic bottlenose dolphins, Tursiops truncatus, in Sarasota, Florida, provide no evidence for alternative male strategies. Canadian Journal of Zoology, 80, Palombit, R. A. (2014). Sexual conflict in nonhuman primates. Advances in the Study of Behavior, 46, Panhuis, T. M., Butlin, R., Zuk, M., and Tregenza, T. (2001). Sexual selection and speciation. Trends in Ecology & Evolution, 16, Patterson, E. M. and Mann, J. (2011). The ecological conditions that favor tool use and innovation in wild bottlenose dolphins (Tursiops sp.). PLoS ONE, 6, e Patterson, E. M. (2012). Ecological and life history factors influence habitat and tool use in wild bottlenose dolphins (Tursiops sp.) (Doctoral dissertation, Georgetown University, Washington DC). Patterson, E. M., Krzyszczyk, E., and Mann, J. (2015). Age-specific foraging performance and reproduction in tool-using wild bottlenose dolphins. Behavioral Ecology doi: /beheco/arv

124 R Development Core Team. (2015). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Randić, S. (2008). Spatial analysis of the distribution and home ranges of male bottlenose dolphins in Shark Bay, Australia (Master's thesis, University of Massachusetts Dartmouth, Dartmouth, MA). Randić, S., Connor, R. C., Sherwin, W. B., and Krützen, M. (2012). A novel mammalian social structure in Indo-Pacific bottlenose dolphins (Tursiops sp.): complex male alliances in an open social network. Proceedings of the Royal Society B, 279, Réale, D., Boussès, P., and Chapuis, J.-L. (1996). Female-biased mortality induced by male sexual harassment in a feral sheep population. Canadian Journal of Zoology, 74, Rodseth, L. and Novak, S. A. (2009). The political significance of gender violence. In M. N. Muller and R. W. Wrangham (Eds.), Sexual Coercion in Primates and Humans (pp ), Cambridge, MA, U.S.A.: Harvard University Press. Rossi, B. H., Nonacs, P., and Pitts-Singer, T. L. (2010). Sexual harassment by males reduces female fecundity in the alfalfa leafcutting bee, Megachile rotundata. Animal Behaviour, 79, Sargeant, B. L., Mann, J., Berggren, P., and Krützen, M. (2005). Specialization and development of beach hunting, a rare foraging behavior, by wild bottlenose dolphins (Tursiops sp.). Canadian Journal of Zoology, 83, Sargeant, B. L., Wirsing, A. J., Heithaus, M. R., and Mann, J. (2007). Can environmental heterogeneity explain individual foraging variation in wild bottlenose dolphins (Tursiops sp.)? Behavioral Ecology and Sociobiology, 61,

125 Sargeant, B. L. and Mann, J. (2009). Developmental evidence for foraging traditions in wild bottlenose dolphins. Animal Behaviour, 78, Schroeder, P.J. (1990). Breeding bottlenose dolphins in captivity. In S. Leatherwood and R. R. Reeves (Eds.), The Bottlenose Dolphin (pp ), New York, NY, U.S.A., Academic Press. Scott, E., Mann, J., and Watson-Capps, J. J. (2005). Aggression in bottlenose dolphins: evidence for sexual coercion, male-male competition, and female tolerance through analysis of tooth-rake marks and behaviour. Behaviour, 142, Smolker, R., Richards, A., Connor, R., and Pepper, J. (1992). Sex differences in patterns of association among Indian Ocean bottlenose dolphins. Behaviour, 123, Smuts, B. B. and Smuts, R. (1993). Male aggression and sexual coercion of females in nonhuman primates and other mammals: evidence and theoretical implications. Advances in the Study of Behavior, 22, Sprogis, K. R., Raudino, H. C., Rankin, R., MacLeod, C. D., and Bejder, L. (2016). Home range size of adult Indo-Pacific bottlenose dolphins (Tursiops aduncus) in a coastal and estuarine system is habitat and sex-specific. Marine Mammal Science, 32, Sundaresan, S. R., Fischhoff, I. R., and Rubenstein, D. (2007). Male harassment influences female movements and associations in Grevy s zebra (Equus grevyi). Behavioral Ecology, 18, Takahashi, Y. and Watanabe, M. (2010). Female reproductive success is affected by selective male harassment in the damselfly Ischnura senegalensis. Animal Behaviour, 79,

126 Tsai, Y. J. and Mann, J. (2013). Dispersal, philopatry, and the role of fission-fusion dynamics in bottlenose dolphins. Marine Mammal Science, 29, Tutin, C. E. G. (1979). Mating patterns and reproductive strategies in a community of wild chimpanzees (Pan troglodytes schweinfurthii). Behavioral Ecology and Sociobiology, Urian, K. W., Hofmann, S. and Wells, R. S. (2009). Fine-scale population structure of bottlenose dolphins (Tursiops truncatus) in Tampa Bay, Florida. Marine Mammal Science, 25, Watson, P. J., Arnqvist, G., and Stallmann, R. R. (1998). Sexual conflict and the energetic costs of mating and mate choice in water striders. The American Naturalist, 151, Watson-Capps, J. J. (2005). Female mating behavior in the context of sexual coercion and female ranging behavior of bottlenose dolphins (Tursiops sp.) in Shark Bay, Western Australia (Doctoral dissertation, Georgetown University, Washington, DC). Watts, D. P. (1998). Coalitionary mate guarding by male chimpanzees at Ngogo, Kibale National Park, Uganda. Behavioral Ecology and Sociobiology, 44, Wearmouth, V. J., Southall, E. J., Morritt, D., Thompson, R. C., Cuthill, I. C., Partridge, J. C., and Sims, D. W. (2012). Year-round sexual harassment as a behavioral mediator of vertebrate population dynamics. Ecological Monographs, 82, Wells, R. (1991). The role of long-term study in understanding the social structure of a bottlenose dolphin community. In K. Pryor and K. S. Norris (Eds.), Dolphin Societies: Discoveries and Puzzles (pp ), Berkeley, CA, U.S.A.: University of California Press. 111

127 Wiszniewski, J., Brown, C., and Möller, L. M. (2012). Complex patterns of male alliance formation in a dolphin social network. Journal of Mammalogy, 93, Würsig, B. and Würsig, M. (1977). The photographic determination of group size, composition, and stability of coastal porpoises (Tursiops truncatus). Science, 198, CHAPTER III Alexander, R. D. (1974). The evolution of social behavior. Annual Review of Ecology and Systematics, 5, Altmann, J. (1974). Observational study of behavior: sampling methods. Behaviour, 49(3), Aplin, L. M., Farine, D. R., Morand-Ferron, J., & Sheldon, B. C. (2012). Social networks predict patch discovery in a wild population of songbirds. Proceedings of the Royal Society of London B: Biological Sciences, rspb Barale, C.L., Rubenstein, D.I., & Beehner, J.C. (2015). Juvenile social relationships reflect adult patterns of behavior in wild geladas. American Journal of Primatology, 77(10), Barocas, A., Ilany, A., Koren, L., Kam, M., & Geffen, E. (2011). Variance in centrality within rock hyrax social networks predicts adult longevity. PLoS One, 6(7), e Berger, V., Lemaître, J. F., Allainé, D., Gaillard, J. M., & Cohas, A. (2015). Early and adult social environments have independent effects on individual fitness in a social vertebrate. Proceedings of the Royal Society B, 282(1813),

128 Bigg, M.A., Olesiuk, P.F., Ellis, G.M., Ford, J.K.B., & Balcomb, K.C. (1990). Social organization and genealogy of resident killer whales (Orcinus orca) in the coastal waters of British Columbia and Washington State. Rep Int Whaling Comm Spec Issue, 12, Brent, L. J. N., Semple, S., Dubuc, C., Heistermann, M., & MacLarnon, A. (2011). Social capital and physiological stress levels in free-ranging adult female rhesus macaques. Physiology & Behavior, 102(1), Brent, L. J., Heilbronner, S. R., Horvath, J. E., Gonzalez-Martinez, J., Ruiz-Lambides, A., Robinson, A. G., Pate Skene, J. H., & Platt, M. L. (2013). Genetic origins of social networks in rhesus macaques. Scientific Reports, 3. Brent, L. J. N. (2015). Friends of friends: Are indirect connections in social networks important to animal behaviour? Animal Behaviour, 103, Cairns, S. J., & Schwager, S. J. (1987). A comparison of association indices. Animal Behaviour, 35(5), Cheney, D. L., Silk, J. B., & Seyfarth, R. M. (2016). Network connections, dyadic bonds and fitness in wild female baboons. Royal Society Open Science, 3(7), Connor, R. C., Smolker, R. A., and Richards, A. F. (1992a). Two levels of alliance formation among male bottlenose dolphins (Tursiops sp.). Proceedings of the National Academy of Sciences of the United States of America, 89(3), Connor, R. C., Smolker, R. A., and Richards, A. F. (1992b). Dolphin alliances and coalitions. In A. H. Harcourt and F. B. M de Waal (Eds.) Coalitions and Alliances in Humans and Other Animals (pp ), New York, NY, U.S.A.: Oxford University Press. 113

129 Connor, R.C., Wells, R.S., Mann, J., & Read, A.J. (2000). The bottlenose dolphin. In: Cetacean Societies (Eds. J. Mann, R.C. Connor, P.L. Tyack, & H. Whitehead), pp The University of Chicago Press: Chicago. Connor, R., Mann, J., & Watson Capps, J. (2006). A sex specific affiliative contact behavior in Indian Ocean bottlenose dolphins, Tursiops sp. Ethology, 112(7), Connor, R. C. and Krützen, M. (2015). Male dolphin alliances in Shark Bay: changing perspectives in a 30-year study. Animal Behaviour, 103, Daisy Kaplan, J., & Connor, R. C. (2007). A preliminary examination of sex differences in tactile interactions among juvenile Atlantic spotted dolphins (Stenella frontalis). Marine Mammal Science, 23(4), Dunbar, R. I. (1998). The social brain hypothesis. Brain, 9(10), Dunbar, R. I. (2009). The social brain hypothesis and its implications for social evolution. Annals of Human Biology, 36(5), Fagen, R. M. (1977). Selection for optimal age-dependent schedules of play behavior. American Naturalist, 111(979), Fairbanks, L. A Reciprocal benefits of allomothering for female vervet monkeys. Animal Behaviour, 40(3), Fairbanks, L. A. (2002). Juvenile vervet monkeys: establishing relationships and practicing skills for the future. Juvenile primates: life history, development and behavior. Chicago: University of Chicago Press. p, Formica, V. A., Wood, C. W., Larsen, W. B., Butterfield, R. E., Augat, M. E., Hougen, H. Y., & Brodie, E. D. (2012). Fitness consequences of social network position in a wild 114

130 population of forked fungus beetles (Bolitotherus cornutus). Journal of Evolutionary Biology, 25(1), Freeberg, T. M., Dunbar, R. I., & Ord, T. J. (2012). Social complexity as a proximate and ultimate factor in communicative complexity. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 367(1597), Frère, C.H., Krützen, M., Mann, J., Watson-Capps, J.J., Tsai, Y.J., Patterson, E.M., Connor, R.C., Bejder, L., and Sherwin, W.B. (2010a). Home range overlap, matrilineal and biparental kinship drive female associations in bottlenose dolphins. Animal Behaviour, 80(3), Frère, C.H., Krützen, M., Mann, J., Connor, R.C., Bejder, L., & Sherwin, W.B. (2010b). Social and genetic interactions drive fitness variation in a free-living dolphin population. Proceedings of the National Academy of Sciences of the United States of America, 107(46), Gibson, Q. A., & Mann, J. (2008a). Early social development in wild bottlenose dolphins: sex differences, individual variation and maternal influence. Animal Behaviour, 76(2), Gibson, Q. A., & Mann, J. (2008b). The size, composition and function of wild bottlenose dolphin (Tursiops sp.) mother calf groups in Shark Bay, Australia. Animal Behaviour, 76(2), Gilby, I. C., Brent, L. J. N., Wroblewski, E. E., Rudicell, R. S., Hahn, B. H., Goodall, J., & Pusey, A. E. (2013). Fitness benefits of coalitionary aggression in male chimpanzees. Behavioral Ecology and Sociobiology, 67(3),

131 Glowacki, L., Isakov, A., Wrangham, R. W., McDermott, R., Fowler, J. H., & Christakis, N. A. (2016). Formation of raiding parties for intergroup violence is mediated by social network structure. Proceedings of the National Academy of Sciences, Gowans, S., Würsig, B., & Karczmarski, L. (2007). The social structure and strategies of delphinids: predictions based on an ecological framework. Advances in Marine Biology, 53, Heithaus, M. R., & Dill, L. M. (2002). Food availability and tiger shark predation risk influence bottlenose dolphin habitat use. Ecology, 83(2), Hrdy, S. B. (1976). Care and exploitation of nonhuman primate infants by conspecifics other than the mother. Advances in the Study of Behavior, 6, Krzyszczyk, E., Patterson, E.M., Stanton, M., & Mann, J. (In prep). The transition to independence: Sex differences in social and behavioural development of wild bottlenose dolphins. Animal Behaviour. Lancaster, J. (1971). Play mothering: The relation between juvenile females and young infants among free-ranging vervet monkeys. Folia Primatologica, 15(3-4), Mann, J. (1999). Behavioral sampling methods for cetaceans: a review and critique. Marine Mammal Science, 15(1), Mann, J., & Smuts, B. B. (1998). Natal attraction: allomaternal care and mother infant separations in wild bottlenose dolphins. Animal Behaviour, 55(5), Mann, J., & Smuts, B. (1999). Behavioral development in wild bottlenose dolphin newborns (Tursiops sp.). Behavior, 136(5),

132 Mann, J., Connor, R.C., Barre, L.M., & Heithaus, M.R. (2000). Female reproductive success in bottlenose dolphins (Tursiops sp.): life history, habitat, provisioning, and group-size effects. Behavioral Ecology, 11(2), Mann, J., & Sargeant, B. (2003). Like mother, like calf: the ontogeny of foraging traditions in wild Indian Ocean bottlenose dolphins (Tursiops sp.). In The Biology of Traditions (Eds. D. M. Fragaszy & S. Perry), pp Cambridge, MA. Cambridge University Press. Mann, J., Sargeant, B. L., Watson-Capps, J. J., Gibson, Q. A., Heithaus, M. R., Connor, R. C., & Patterson, E. (2008). Why do dolphins carry sponges? PLoS One, 3(12), e3868. Mann, J., Stanton, M. A., Patterson, E. M., Bienenstock, E. J., and Singh, L. O. (2012). Social networks reveal cultural behaviour in tool-using dolphins. Nature Communications, 3, 980. Mariette, M. M., Cathaud, C., Chambon, R., & Vignal, C. (2013). Juvenile social experience affects pairing success at adulthood: congruence with the loser effect? Proceedings of the Royal Society of London B: Biological Sciences, 280(1767), McDonald, D. B. (2007). Predicting fate from early connectivity in a social network. Proceedings of the National Academy of Sciences, 104(26), McHugh, K. A., Allen, J. B., Barleycorn, A. A., & Wells, R. S. (2011). Natal philopatry, ranging behavior, and habitat selection of juvenile bottlenose dolphins in Sarasota Bay, Florida. Journal of Mammalogy, 92(6), Murray, C. M., Lonsdorf, E. V., Stanton, M. A., Wellens, K. R., Miller, J. A., Goodall, J., & Pusey, A. E. (2014). Early social exposure in wild chimpanzees: Mothers with sons are more gregarious than mothers with daughters. Proceedings of the National Academy of Sciences, 111(51),

133 Nuñez, C. M., Adelman, J. S., & Rubenstein, D. I. (2015). Sociality increases juvenile survival after a catastrophic event in the feral horse (Equus caballus). Behavioral Ecology, 26(1), Oh, K. P., & Badyaev, A. V. (2010). Structure of social networks in a passerine bird: consequences for sexual selection and the evolution of mating strategies. American Naturalist, 176(3), E80 E89. Parrish, J. K., & Edelstein-Keshet, L. (1999). Complexity, pattern, and evolutionary trade-offs in animal aggregation. Science, 284(5411), Patterson, E. M., & Mann, J. (2011). The ecological conditions that favor tool use and innovation in wild bottlenose dolphins (Tursiops sp.). PLoS One, 6(7), e Pereira, M. E. (1988). Effects of age and sex on intra-group spacing behaviour in juvenile savannah baboons, Papio cynocephalus cynocephalus. Animal Behaviour, 36(1), Pereira, M. E., & Fairbanks, L. A. (Eds). (2002). Juvenile primates: Life history, development, and behavior. New York: University of Chicago Press. R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL Ryder, T. B., McDonald, D. B., Blake, J. G., Parker, P. G., & Loiselle, B. A. (2008). Social networks in the lek-mating wire-tailed manakin (Pipra filicauda). Proceedings of the Royal Society B: Biological Sciences, 275(1641), Ryder, T. B., Parker, P. G., Blake, J. G., & Loiselle, B. A. (2009). It takes two to tango: reproductive skew and social correlates of male mating success in a lek-breeding bird. 118

134 Proceedings of the Royal Society of London B: Biological Sciences, 276(1666), Sargeant, B. L., Wirsing, A. J., Heithaus, M. R., & Mann, J. (2007). Can environmental heterogeneity explain individual foraging variation in wild bottlenose dolphins (Tursiops sp.)? Behavioral Ecology and Sociobiology, 61(5), Sargeant, B. L. and Mann, J. (2009). Developmental evidence for foraging traditions in wild bottlenose dolphins. Animal Behaviour, 78(3), Scott, E., Mann, J., & Watson-Capps, J. J. (2005). Aggression in bottlenose dolphins: evidence for sexual coercion, male-male competition, and female tolerance through analysis of tooth-rake marks and behaviour. Behaviour, 142(1), Seyfarth, R. M., & Cheney, D. L. (2012). The evolutionary origins of friendship. Annual Review of Psychology, 63, Silk, J. B. (1999). Why are infants so attractive to others? The form and function of infant handling in bonnet macaques. Animal Behaviour, 57(5), Silk, J. B. (2002). The form and function of reconciliation in primates. Annual Review of Anthropology, Silk, J. B. (2007a). Social components of fitness in primate groups. Science, 317(5843), Silk, J. B. (2007b). The adaptive value of sociality in mammalian groups. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 362(1480),

135 Silk, J. B., Rendall, D., Cheney, D. L., & Seyfarth, R. M. (2003a). Natal attraction in adult female baboons (Papio cynocephalus ursinus) in the Moremi Reserve, Botswana. Ethology, 109(8), Silk, J. B., Alberts, S. C., & Altmann, J. (2003b). Social bonds of female baboons enhance infant survival. Science, 302(5648), Silk, J. B., Altmann, J., & Alberts, S. C. (2006a). Social relationships among adult female baboons (Papio cynocephalus) I. Variation in the strength of social bonds. Behavioral Ecology and Sociobiology, 61(2), Silk, J. B., Alberts, S. C., & Altmann, J. (2006b). Social relationships among adult female baboons (Papio cynocephalus) II. Variation in the quality and stability of social bonds. Behavioral Ecology and Sociobiology, 61(2), Silk, J. B., Beehner, J. C., Bergman, T. J., Crockford, C., Engh, A. L., Moscovice, L. R., Wittig, R. M., Seyfarth, R. M., & Cheney, D. L. (2009). The benefits of social capital: close social bonds among female baboons enhance offspring survival. Proceedings of the Royal Society B: Biological Sciences, 276(1670), Silk, J. B., Beehner, J. C., Bergman, T. J., Crockford, C., Engh, A. L., Moscovice, L. R., Wittig, R. M., Seyfarth, R. M., & Cheney, D. L. (2010). Strong and consistent social bonds enhance the longevity of female baboons. Current Biology, 20(15), Smolker, R. A., Richards, A. F., Connor, R. C., & Pepper, J. W. (1992). Sex differences in patterns of association among Indian Ocean bottlenose dolphins. Behavior, 123(1),

136 Smuts, B. B., & Smuts, R. W. (1993). Male aggression and sexual coercion of females in nonhuman primates and other mammals: evidence and theoretical implications. Advances in the Study of Behavior, 22(22), Spear, L. P. (2000). The adolescent brain and age-related behavioral manifestations. Neuroscience & Biobehavioral Reviews, 24(4), Stanton, M.A., & Mann, J. (2012). Early social networks predict survival in wild bottlenose dolphins. PloS One, 7(10), e Stanton, M.A., Gibson, Q.A., & Mann, J. (2011). When mum s away: a study of mother and calf ego networks during separations in wild bottlenose dolphins (Tursiops sp.). Animal Behaviour, 82(2), Sterck, E. H., Watts, D. P., & van Schaik, C. P. (1997). The evolution of female social relationships in nonhuman primates. Behavioral Ecology and Sociobiology, 41(5), Tsai, Y. J., & Mann, J. (2013). Dispersal, philopatry, and the role of fission-fusion dynamics in bottlenose dolphins. Marine Mammal Science, 29(2), Tung, J., Archie, E. A., Altmann, J., & Alberts, S. C. (2016). Cumulative early life adversity predicts longevity in wild baboons. Nature Communications, 7, Vander Wal, E., Festa-Bianchet, M., Reale, D., Coltman, D. W., & Pelletier, F. (2015). Sexbased differences in the adaptive value of social behavior contrasted against morphology and environment. Ecology, 96(3), de Waal, F. B. and Tyack, P. L. (Eds.). (2009). Animal social complexity: intelligence, culture, and individualized societies. Harvard University Press. 121

137 Wallen, M. M., Patterson, E. M., Krzyszczyk, E., & Mann, J. (2016). The ecological costs to females in a system with allied sexual coercion. Animal Behaviour, 115, Weber, N., Carter, S. P., Dall, S. R., Delahay, R. J., McDonald, J. L., Bearhop, S., & McDonald, R. A. (2013). Badger social networks correlate with tuberculosis infection. Current Biology, 23(20), R915-R916. Wey, T. W., & Blumstein, D. T. (2012). Social attributes and associated performance measures in marmots: bigger male bullies and weakly affiliating females have higher annual reproductive success. Behavioral Ecology and Sociobiology, 66(7), Wey, T. W., Burger, J. R., Ebensperger, L. A., & Hayes, L. D. (2013). Reproductive correlates of social network variation in plurally breeding degus (Octodon degus). Animal Behaviour, 85(6), Whitehead, H., Christal, J., & Tyack, P.L. (2000). Studying cetacean social structure in space and time. In: Cetacean Societies (Eds. J. Mann, R.C. Connor, P.L. Tyack, & H. Whitehead), pp The University of Chicago Press: Chicago. Whitehead, H. (2008). Analyzing animal societies: quantitative methods for vertebrate social analysis. University of Chicago Press. Wittemyer, G., Douglas-Hamilton, I., and Getz, W. M. (2005). The socioecology of elephants: analysis of the processes creating multitiered social structures. Animal Behaviour, 69(6), Wrangham, R. W. (1980). An ecological model of female-bonded primate groups. Behaviour, 75(3),

138 APPENDIX A Chapter II Supplemental Material METHODS Single male associations As noted in the Methods, allied male aggression directed towards females is commonly observed, yet aggression by a lone male towards a female has never been observed. As evident from Table I (in the main text), a single adult male is rarely found with one or more females. Therefore, it is unlikely that a lone male can coerce a female in a mating context. To support the argument that coercion is acting only when multiple males are present, here we examined survey sightings in which a single male was present with at least one adult female, and compared the same metrics of female ranging and habitat use to surveys when no males and >1 male was present. A subset of 11 adult females had enough data to include surveys where a single male was present. The average number of surveys (± SD) per female, per factor level for all ranging analyses was ± (not cycling, males absent), ± (not cycling, one male present), ± (not cycling, >1 male present), ± (cycling, males absent), ± (cycling, one male present), and ± (cycling, >1 male present). The average number of surveys (± SD) per female, per factor level for all habitat analyses was ± (not cycling, males absent), ± (not cycling, one male present), ± 123

139 14.92 (not cycling, >1 male present), ± (cycling, males absent), ± (cycling, one male present), and ± (cycling, >1 male present). RESULTS Ranging There was no interaction between cycling status and male presence (F = 0.16, P = 0.1) on female centroid distance from baseline. Only male presence (maxt = 4.5, P < ) and not cycling status (Z = -1.28, P > 0.84) influenced the distance (natural log-transformed) a female was found from her baseline centroid (Fig. S1a). Both cycling females and non-cycling females were found farthest from their baseline centroid when >1 male was present, compared to when males were absent (non-cycling: Z = -3.23, P < ; cycling: Z = -3.54, P < ), while baseline distance when a single male was present did not differ from either the >1 male distance, nor the male absent distance (Fig. S1a). There was also no interaction between cycling status and male presence (F = 0.18, P = 0.84) on female average ranging distance. Cycling status did not affect female average ranging distance (Z = -0.15, P = 0.88), but male presence was marginally significant (maxt = 2.17, P = 0.06), so we continued with a priori contrasts (Fig. S1b). Surveys in which only a single male was present did not differ from surveys in which males were absent (non-cycling: Z = 1.16, P = 0.27; cycling: Z = 0.82, P = 0.41) or >1 male was present (non-cycling: Z = 0.72, P = 0.48; cycling: Z = -0.99, P = 0.33). As in the full analyses reported in the main text, female average 124

140 ranging distance in the subset remained much lower in surveys where males were absent compared to those where >1 male was present (non-cycling: Z = -2.07, P = 0.02; cycling: Z = , P = 0.04), but this did not remain significant after a Bonferroni correction (Fig. S1b). Habitat use Overall habitat use (Table S1) and primary habitat use (Fig. S1c) did not differ from baseline when females were with single males during cycling and non-cycling states. DISCUSSION Results from analyzing the surveys in which only a single male was present are largely consistent with those presented in the full text. However, given the increased variance that comes with a much reduced sample size, several of our findings were either only marginally significant, or became insignificant after Bonferroni adjustments. Nevertheless, generally speaking the presence of a single male did not alter female ranging or habitat use, but the presence of more than one male did. This further supports the hypothesis that females experience ecological costs from sexually coercive allied males. Somewhat surprisingly, the single-male associations appeared to be intermediate between no-male and more-than-one-male associations. Such a result could be explained by the fact that single-male associations actually confound two distinct contexts that likely differ in their effect on female space use. Some of the single-male associations likely present no threat to females if the lone male is not in an alliance. Approximately one-third of adult males in the Shark Bay population are not in alliances, defined 125

141 as having no close association (>0.5 half-weight COA) with another male (unpublished data). On the other hand, some single-male associations likely occur when alliance partners are nearby but not in the group (>10 m away), where coercion is possible. Although technically not in association, allied males would still be close enough to assist each other should the female attempt to escape. While surveys are not ideal for differentiating between these contexts, analysis of focal follows of individual females may help elucidate the nature of single male, single female associations. 126

142 FIGURES AND TABLES Distance from baseline centroid (km) (a) a ab b Male presence 0 males 1 male >1 male a ab b Average ranging distance (km) (b) Primary habitat selection ratio (c) Not cycling Cycling 127

143 Figure S1. Adult female (N = 11) (a) distance from the baseline centroid, (b) average ranging distance, and (c) primary habitat selection ratios for the subset female sample analysed with a single male present for each combination of cycling and male presence factor levels. Error bars indicate ±1 SE. Letters indicate significant differences after Bonferroni adjustments. Distances in (a) were natural log-transformed for analysis but raw distances are shown for illustrative purposes. Table S1. MANOVA results for overall female habitat use selection ratios and the effect of female cycling status and male presence. df = degrees of freedom, MS = mean squares, F = model statistic from the permuted F-tests, R 2 = adjusted R-squared or effect size, P = significance value. df MS F R 2 P Females Cycling (N = 11) Male presence Cycling*Male presence Residual

144 APPENDIX B Publishing Note Chapter II was published in the journal Animal Behaviour. This journal does not require that authors receive permission prior to publishing the authors articles in their own theses ( 129

145 APPENDIX C Chapter III Supplemental Material FIGURES AND TABLES Figure S1. Each principal component on the x-axis (1-5) with the cumulative variance explained (area under the line). 130

146 Figure S2. Plot of each of the 33 juvenile females in their relative locations in the principal component space. 131

147 Figure S3. Plot of each social network metric relative to its weight on each principal component. 132