|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 STRESS AND AFFILIATION AMONG WILD FEMALE PRIMATES: EFFECTS OF GROUP SIZE, RISK, AND REPRODUCTIVE CONDITION IN A DYNAMIC FOREST COMMUNITY By ERIN E. EHMKE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010
2 2010 Erin E. Ehmke
3 To Jim
4 ACKNOWLEDGMENTS Funding for this research was provided by the National Science Foundation (SBR9722840; BCS 0078967; BCS 035 2316) and the L.S.B. Leakey Foundation awarded to S. Boinski. I thank the University of Fl orida for additional financial support during various stages of my graduate career, incl uding the Departments of Anthropology and Zoology teacher assistantships and research assistantships. To Toni Ziegler and Dan Wittwer at the Wisconsin Regional Primate Research Center, I ex press deep gratitude for not only their assay services, but for their support and patient correspondence regarding nearly every aspec t of the hormone collection and analys is process. I wish to thank the many people who have given me support and encouragement throughout the Ph.D. process. First, and most importantly, I would like to acknowledge the support and guidance of my committee members, Sue Boinski (chair), David Daegling, John Krigbaum and St even Phelps. My relationshi p with Bo extends back to my undergraduate years, as a student in her Primate Behavior course and as a volunteer data princess. He r energetic genius is inspiring, and her confidence in me facilitated my path to study wild monkeysa dream that I was not quite sure how to make a reality. I owe Bo many thanks for her support and encouragement over the years, for giving me the opportunity to work at her field site, and fo r her devotion to my success as a graduate student and beyond. I am a stronger person because of her. I would like to thank David Daeg ling, the consummate profe ssional, for contributing a sense of sanity and stability to my graduate ex perience. As a member of both my M.A. and Ph.D. committees, he has always been the one I could turn to for sound advice. I thank John Krigbaum for being my cheerleader I am not sure how I would have made it through were it not for his encouragement, humor, and suppor t. My dissertation topic
5 may have never developed as it did were it not for Steven Phelps I thank him for introducing me to the world of behavioral endoc rinology and for helping me to feel like I am not completely in over my head. Additional thanks are due to Allan Bu rns, Ken Sassaman, Karen Jones, Pat Gaither King, Rhonda Riley and Juanita Bagna ll in the Anthropology Department office at The University of Florida. Their extraordinary behind-the-scenes guidance, sage knowledge of university require ments, and talent for all-th ings-paperwork helped to make the graduate school process much more manageable than should have ever been expected. When dealing with fr azzled grad students all day, it is a wonder how they always managed to have a smile on their faces and calming words of advice. They are truly the unsung heroes of graduate school. I wish to thank STINASU (Suriname Natu re Conservation Foundation), for without their essential support, our field research in Suriname would not have been possible. The valuable field assistance of John A nderson, Claire Sheller, Andy Dosmann, Miranda Davis, Tory Shelley, Laurie Kauffman, Liz Fly and Julie Massicotte is much appreciated. Without them, my dissertation research would not have been possible, and my return to Raleighvallen would not have been nearly as fun. I will forever be proud to be one of the Boinski uma! I thank my friends and family who hav e stood by me through the years and who patiently listened while I vented about the la test trials and tribulations of graduate school. I especially thank Jackson Frechette, Carson Phillips, Carrie Vath and Jon Colburn for taking the br unt of such ranting sessions. I appreciate the friendship of all of my peers, and that of Anna Vi ck, Chad Maxwell, and Jennifer Hotzman in particular
6 my graduate experience would not have been the same without them. I thank Kimberly Skipper for being my long-time friend and for being patient, understanding and supportive through my years as a grad school recluse. I will always cherish the friendship of Michelle Herndon a nd the inspiration that she is in my life. Every day is dedicated to what she would hav e brought to the world, and I am a better person for it. I am grateful beyond words to Laurie K auffman for being my friend, colleague and support system throughout these past eight (y es, eight!) years. She has been my partner in almost everything I have experienc ed since 2002, and as excited as I am that we are now moving on to bigger and brighter fu tures, I am sad (and a little nervous) at the realization that we will no longer be connect ed at the hip. I look forward, however, to many, many years of collaborat ion and to crazy conference reunions. I thank Kari Bagnall for introducing me to pr imates, for trusting me to care for her sanctuary, and for believing in my ability to one day make the world a better place for wild and captive animals. My heart will always be with the monkeys of Jungle Friends Primate Sanctuary, the true in spiration for my research, an d I thank them for allowing me to be a part of their lives. It goes without saying that I owe everything to my parents. I thank them for allowing me to become the person I am today for letting me go to forge my own path but always being there in case I took a wrong turn. The confidence to follow my dreams is solely attributable to the unwavering lo ve and support from my family, who may not always understand my choices in life (e.g., to leave the comfor ts of home and live in the bug-infested, snake-ridden rainforest), but never question them. I thank Reeses and
7 Gizmo, for without their unconditional love and cuddles, I never would have made it through many stressful all-nighters. My world would never be what it is, and I w ould never be what I am, were it not for my husband, Jim. I could never put into wo rds what he means to me and how thankful I am that he is in my life. He is my rock. He makes me laugh and lets me cry, and that has been invaluable during these past few years. It has not been easy, for either of us, to start a life together while dealing with the stressors of grad school, but we made it through. He supported me in every way possi ble, much more so than I could have ever hoped for. I thank Jim for hi s love, his support, and especially for his patience with this dissertation that took over our lives. We can now move on, and I am excited beyond words to discover what adventures await us. And of course, I thank the c apuchins of Raleighvallen. The two years of my life that I spent with them will forever be my most memorable. I am especially grateful to Gina, Banana, Jane, Little Horn s, Carol, Kate, Tinkerbell, Luna, Mrs. Slocum, and Miss Brahms for patiently humoring me in my obsession over their poo.
8 TABLE OF CONTENTS page ACKNOWLEDG MENTS .................................................................................................. 4LIST OF TABLES .......................................................................................................... 13LIST OF FIGURES ........................................................................................................ 15LIST OF ABBR EVIATIONS ........................................................................................... 18ABSTRACT ................................................................................................................... 19 CHA PTER 1 INTRODUCTION AND LI TERATURE REVIEW ..................................................... 21General Research Goal and Specific Ob jectives .................................................... 21Literature Review .................................................................................................... 22The Physiology of Stre ss and Its Meas urement ............................................... 22Primate Sociality, Competition, and Social St ructure: An Historic Perspective ................................................................................................... 25Mammalian Social Bo nds: A Re view ................................................................ 26Mammalian Social Bonds: A Evolutionary Pe rspective .................................... 29The Biological Ba sis of Gr oomi ng .................................................................... 30The Social Functi ons of Gr oomin g ................................................................... 33Cebus : Ecologically Diverse, Socially Similar? ................................................. 35Dissertation Overview: What Lies Ahead ................................................................ 372 THE EFFECT OF GROUP SIZE ON TH E SOCIAL AND ECOLOGICAL MIL IEU OF FEMALE BROW N CAPUCHI NS ....................................................................... 41Introduc tion ............................................................................................................. 41Hypothes es ............................................................................................................. 42Daily Travel Distance Increases with Group Size ............................................. 42Individual Foraging Time in a Preferred, Monopol izable Food Tree Decreases with Incr eased Group Size .......................................................... 42Predator Alarm Frequency (per indivi dual) is Inversely Associated with Group Si ze .................................................................................................... 43Rates of Female Agonism Increase with Group Size and Group Size Affects the Strength of Female Dominance Hierarchies ............................................ 43Rates of Female Grooming Increase with Gr oup Size ..................................... 44Female Brown Capuchins in a Lar ge Troop Demonstrate More Seasonal Variation in Patterns of Grooming and Agonism than Do the Females in a Small Tr oop ................................................................................................... 44Methodology ........................................................................................................... 45
9 Site Descr iption ................................................................................................ 45Study An imal .................................................................................................... 46Data Coll ection ................................................................................................. 47Data Anal ysis ................................................................................................... 48Behavioral rates ......................................................................................... 48Dominance hi erarchy ................................................................................. 49Statistical analysis ...................................................................................... 50Results .................................................................................................................... 50Group Size, Home R ange and Habita t Use ...................................................... 50Group Size, Density and Variat ion in Travel Distance ...................................... 52Group Size and Variation in Maximiliana maripa Feeding ................................ 52Group Size and Variati on in Ala rm Calls .......................................................... 53Group Size and Variation in Female Agonism .................................................. 55Group Size and Variation in Female Grooming ................................................ 57Group Size and Seasonal Variat ion in Social Behavior .................................... 58Discuss ion .............................................................................................................. 59Daily Travel Distance Increases with Group Size ............................................. 59Individual Foraging Time in a Preferred, Monopol izable Food Tree ( Maximiliana maripa ) Decreases with Incr eased Group Size ........................ 61Predator Alarm Frequency (per indivi dual) is Inversely Associated with Group Si ze .................................................................................................... 61Rates of Female Agonism Increase with Group Size and Group Size Affects the Strength of Female Dominance Hierarchies ............................................ 64Rates of Female Grooming Increase with Gr oup Size ..................................... 65Female Brown Capuchins in a Lar ger Troop Demonstrate More Seasonal Variation in Patterns of Grooming and Agonism than Do the Females in a Smaller Troop ................................................................................................ 66Conclusi on .............................................................................................................. 673 DISPARATE TROOP SIZE AND STRESS: THE EFF ECTS OF A HIGH RISK, LOW ENERGY ENVIRONMENT ON FEMALE BROWN CAPUCHINS .................. 83Introduc tion ............................................................................................................. 83Hypothes es ............................................................................................................. 84Group Size Affects the Level of Stress Females Incur Due to Predation Risk .. 84Group Size Affects the Levels of Stress Females Incur During Food Shortages ...................................................................................................... 85Rank-Related Stress is a Factor of Group Si ze ................................................ 86Methodology ........................................................................................................... 87Site Descr iption ................................................................................................ 87Study An imal .................................................................................................... 88Data Coll ection ................................................................................................. 89Hormonal Sampling and Analysis ..................................................................... 89Statistical Analysis ............................................................................................ 90Results .................................................................................................................... 91Fecal Cortisol Profiles ...................................................................................... 91Diurnal Variation in Fe male Cortis ol Levels ...................................................... 91
10 Troop Size, Female Cortisol Levels and Pred ation Risk .................................. 92Troop Size, Stress, and Seasona lity ................................................................ 93Troop Size, Stress, and Social Rank ................................................................ 94Discuss ion .............................................................................................................. 95Group Size Affects the Level of Stress Females Incur Due to Predation Risk .. 95Group Size Affects the Level of Stress Females Incur During Food Shortages ...................................................................................................... 96Rank-Related Stress is a Factor of Group Si ze ................................................ 99Conclusi on ............................................................................................................ 1004 FEMALE-BONDED PRIMATES? GROO MING AND PROXIMITY PATT ERNS OF FEMALE BROWN CAP UCHINS IN SU RINAME ............................................ 108Introduc tion ........................................................................................................... 108Using Affiliative Behavior as a Proxy Measure of Social Bonds ..................... 108A Female-Bonde d Prim ate? ........................................................................... 109Hypothes es ........................................................................................................... 111Social Bonds among Female Brown Capu chins are Variable Over Time ....... 111Dyadic Female Relationships Vary in Response to Ecological Pressures...... 112Relationships among Females Strengthen wi th the Presence of an Infant .... 114Methodology ......................................................................................................... 114Site Descr iption .............................................................................................. 114Study Animal .................................................................................................. 115Data Coll ection ............................................................................................... 116Behavioral De finitions ..................................................................................... 117Data Anal ysis ................................................................................................. 118Dominance ranks ..................................................................................... 118Proximity scores ....................................................................................... 119Grooming ................................................................................................. 120Statistical analysis .................................................................................... 120Results .................................................................................................................. 121Dominance Hie rarchy ..................................................................................... 121Female Coa litions ........................................................................................... 122Troop Patterns of Affiliati on ............................................................................ 123Proximity .................................................................................................. 123Grooming ................................................................................................. 123Individual Patterns of Affilia tion ...................................................................... 124Proximity .................................................................................................. 124Grooming ................................................................................................. 125Grooming versus proximity ...................................................................... 126Monthly Variation in Aff iliation ........................................................................ 127Proximity .................................................................................................. 127Grooming ................................................................................................. 127Seasonal Variation in Affilia tion ...................................................................... 129Proximity .................................................................................................. 129Grooming ................................................................................................. 129The Effect of Infants on Female A ffiliati on ...................................................... 131
11 Proximity .................................................................................................. 131Grooming ................................................................................................. 133Discuss ion ............................................................................................................ 134Outcome of Hypothes es ................................................................................. 136Social bonds between adult female brown capuchins are variable over time ....................................................................................................... 136Dyadic female relationships vary in response to ecological pressures .... 137Relationships among females strengthen with the presence of an infant 139Female-Male Re lationships ............................................................................ 141Associations between females with infants and adult males .................... 142Associations between reproductively-cycling females and adult males ... 143Conclusi on ............................................................................................................ 1455 DO FEMALE BROWN CAPUCHIN MO NKEYS US E AFFILIATIVE BEHAVIOR TO MEDIATE STRESS? ....................................................................................... 169Introduc tion ........................................................................................................... 169Stress and So ciality ........................................................................................ 170Stress Hormones and Female Reproductive Condition .................................. 173Pregnancy and la ctation .......................................................................... 173Estrus cycling ........................................................................................... 175Hypothes es ........................................................................................................... 176Methodology ......................................................................................................... 177Site Descr iption .............................................................................................. 177Study Animal .................................................................................................. 178Data Coll ection ............................................................................................... 179Behavioral and Reproduc tive Defi nitions ........................................................ 179Hormonal Sampling and Analysis ................................................................... 181Statistical Analysis .......................................................................................... 182Results .................................................................................................................. 182Association between Social Proxim ity and Female Co rtisol Levels ................ 182Association between Grooming (given) and Female Cort isol Levels .............. 183Association between Grooming (received) and Female Cortisol Levels ......... 184Reproductive Condition, Female Cortisol Levels and Affiliative Behavior ...... 185Cortisol levels ........................................................................................... 185Proximity .................................................................................................. 186Grooming (g iven) ..................................................................................... 187Grooming (rec eived) ................................................................................ 189Discuss ion ............................................................................................................ 189Social Affiliation and Cortisol levels ................................................................ 189The Influence of Reproductive C ondition on Cort isol Levels .......................... 193The Influence of Reproductive Condi tion on Affilia tive Behavior .................... 195Conclusi on ............................................................................................................ 1986 CONCLUS ION ...................................................................................................... 209
12 APPENDIX A CHAPTER 2 RESU LTS SUMMA RY ..................................................................... 214B ADDITIONAL RESULTS PERTAINING TO TABL ES 4-5 AND 4-6: DYADIC PATTERNS OF FEMALE AFFI LIATION (TROOP A) ........................................... 217Proxim ity ............................................................................................................... 217Grooming .............................................................................................................. 219C TOTAL PROXIMITY SCORES (PS) FOR ALL FEMALE-FEMALE AND FEMALE-MALE DYADS (TROOP A) (JANUARY-DECEM BER, 2006) ................ 221D TOTAL GROOMING DURATIONS FO R ALL FEMAL E-FEMALE AND FEMALE-MALE DYADS (TROOP A) (JANUARY-DECEM BER, 2006) ................ 223E GRAPHICAL REPRESENTATION OF EACH FEMALES BASELINE CORTISOL VALUE A ND MONTHLY VARIATION ABOUT THAT MEAN (JANUARY-NO VEMBER, 2006) ........................................................................... 226LIST OF RE FERENCES ............................................................................................. 230BIOGRAPHICAL SKETCH .......................................................................................... 254
13 LIST OF TABLES Table page 1-1 Socioecological comparison of Cebus populations at vari ous field sites. ........... 391-2 Overview of dissertat ion chapters 2 to 5. ............................................................ 402-1 Reported (or calculated mean) group size of C. apella troops at various sites. .. 702-2 Composition of the study groups, Troop A and Troop B, from January to December 2006.. ................................................................................................ 702-3 Operational defin itions for female C. apella agonistic and grooming behavior, recorded ad libitum ............................................................................................ 712-4 Summary of the observed intragroup agonistic bouts involving females (with determinable winner and lose r) for Troop A and Troop B.. .............................. 722-5 Troop A female dominance matrix based on the outcomes of dyadic agonism ............................................................................................................. 722-6 Differential benefits and costs for indi vidual female brown capuchins residing in a large group (Troop A) versus a small group (Troop B) in Raleighvallen, Suriname. ......................................................................................................... 733-1 Baseline values of individual female fecal cortisol concentrati ons.. .................. 1024-1 Summary of the dominance rank, approximate age, and reproductive history of the adult females in Troop A, as well as general life history notes for the adult male s. ...................................................................................................... 1474-2 Female dominance matrix depicting a (weakly) linear hierarchy based on the outcomes of dyadic agonism, transitivity, and anecdotal evidence. .................. 1484-3 Average monthly proximity score s ( SE) of Troop A females.. ........................ 1484-4 Frequency and duration of female groom ing dyads (by age/sex class), from January to Dece mber 2006.. ............................................................................ 1494-5 Total proximity scores for adult dyads, January to December 2006.. ............... 1504-6 Total grooming duration (min) for adul t dyads, January to December 2006.. ... 1514-7 Monthly variation in females primary adult proximity partners, based on the highest dyadic proximity score/month (shown).. ............................................... 152
14 4-8 Monthly variation in proximity (p) and grooming (g) strong bonds, comprising the top 10% of monthly pr oximity scores and grooming durations (such that the focal female is the groomer, dyadic partner the recipient) among adult fema le dy ads.. ............................................................................. 1534-9 Monthly variation in proximity (p) and grooming (g) strong bonds, comprising the top 10% of monthly pr oximity scores and grooming durations among adult dyads.. ......................................................................................... 1554-10 Monthly variation in females pr imary adult grooming partner, based on the highest dyadic grooming rate (m in/h) per mont h (shown ).. ............................... 1574-11 Females primary adult partner per month (proximity /grooming).. .................... 1585-1 Summary of the dominance rank, approximate age, birthing history and reproductive classification (for pur poses of my study) of the adult C. apella focal females. ................................................................................................... 2005-2 Females mean monthly values ( SE) of cortisol (F), to tal proximity score (TPS), total female proximity score ( TFPS) and male proximity score (MPS).. 2005-3 Range of values for each females X monthly cortisol levels and total monthly proximity to females (TFPS) and to ma les (MPS).. ............................. 2015-4 Association between a females frequen cy of occupying a central position in the troop and her social proximity to adult females (TFPS) and adult males (MPS).. ............................................................................................................. 2015-5 Relationship between months with and without proxim ity strong bonds andX monthly cortisol (F) levels (ng/g).. ..................................................................... 2025-6 Relationship between months with and without grooming (given) strong bonds andX monthly cortisol (F) levels (ng/g).. ................................................ 2025-7 Relationship between months with and without grooming (received) strong bonds andX monthly cortisol (F) levels (ng/g).. ................................................ 2035-8 Effect of infant age on mothers ( n = 4) monthly proximity to adult females (TFPS) ( n = 6) and adult males (MPS) ( n = 3).. ................................................ 2035-9 Monthly association between gestation length, proximity to females (TFPS) and proximity to males (MPS).. ......................................................................... 204
15 LIST OF FIGURES Figure page 2-1 Habitat composition of the home ranges of Troop A and Tr oop B. ..................... 742-2 Comparison of overall time (%) that Troop A and Troop B spent in each habitat ty pe. ........................................................................................................ 752-3 Seasonal comparison of Tr oop As habitat use (% time). .................................. 762-4 Seasonal comparison of Tr oop Bs habitat use (% time). .................................. 762-5 Seasonal comparison of average canopy cover (0-3 incremental scale that increases with vegetation densit y) for Troop A and Troop B. ............................. 772-6 Average canopy cover by habita t type for Troop A and Troop B. ....................... 772-7 Percentage of time Troops A and B re mained stationary ( vs. travel) during the morning (7:00 to 11:00 a.m.), mid-day (11:00 to 2:00), and afternoon (2:00 to 5:00 p.m.) hour s. ................................................................................... 782-8 Seasonal comparison of daily travel rates (m/h SE) for Troops A and B.. ....... 782-9 Monthly group and individual rates (ala rms/h SE) of aerial alarms (AA) and terrestrial predator alarms (TPA) for Troops A and B. ........................................ 792-10 Seasonal effect on group rates of aerial alarms ( AAs/h SE) for Troops A (n = 23) and B ( n = 9). ............................................................................................ 802-11 Seasonal effect on the rates of aerial alarms (AAs/h SE) for individuals in Troop A and B. ................................................................................................... 802-12 Seasonal effect on the group rates of terrestrial predat or alarms (TPAs) (TPAs/h SE) for Troops A ( n = 23) and B ( n = 9). ............................................ 812-13 Seasonal effect on the rates of terr estrial predator alarms (TPAs/h SE) for individuals in Troop A and B.. ............................................................................. 812-14 Controlling for group size, the mont hly rate (bouts/h SE) at which the average female in Troop A and Troop B experienced agonism (as actor vs. recipient) with troop-mate s. ................................................................................ 822-15 Monthly rate (min/h SE) that th e average female in Troop A and Troop B acted as the groomer and as t he recipient duri ng grooming.. ............................. 823-1 Troop-level fecal cortisol (ng/g SE) profiles of Group A and Group B females, based on average m onthly baselin e values. ...................................... 103
16 3-2 Morning versus afternoon cortisol le vels (ng/g SE) for Troop A and Troop B females. ............................................................................................................ 1033-3 Monthly comparison of t he rate of aerial alarms ( AA/h) produced by Troop A, observed number of Harpy Eagle attacks and/or attendance to the troop within 50 m, and mean cortisol le vels of Troop A females. ............................... 1043-4 Seasonal comparison of average basel ine cortisol levels (ng/g SE) for Troop A and B fe males. ................................................................................... 1053-5 Mean baseline cortisol concentrations ( ng/g SD) of individual adult females in Troop A by their respective dominance rank (i n parenthes es). ..................... 1063-6 Seasonal cortisol concentrations ( ng/g SE) of Troop A females and the rate at which females were the recipients of social agonism from conspecifics (bouts/ h). .......................................................................................................... 1074-1 Deconstructing intra-sexual social bonds: Various usages of the term social bond and its synonyms, as found in the primate literature. .............................. 1594-2 Total duration (min) females gr oomed and received grooming from adult male and female conspecifics. .......................................................................... 1604-3 The monthly rate (min/h) at whic h females groomed up the hierarchy as opposed to down the hierarchy. ........................................................................ 1614-4 Total duration (min) that adult fe males groomed troop-mates grouped by age/sex class (black bars), and the total duration that adult females received grooming from troop-mates grouped by age/sex classes (grey bars). .............. 1624-5 Total duration and average bout length of and grooming. Females groomed males (grey bars) for a longer total duration than females groomed females (black bars) ........................................................................................ 1634-6 Association between total proximit y scores and grooming durations (min), including all female-female and female-male dyads ( rs = 0.41, P = 0.0010, n = 63) ................................................................................................................ 1644-7 Seasonal variation in female proximity to females (black bars) and to males (grey bar s). ....................................................................................................... 1654-8 Seasonal variation in the rate at which females groomed females (black bars) and females groomed males (grey bars). ......................................................... 1654-9 The monthly rate (min/h) at whic h females in each season groomed up the hierarchy (black bars) as opposed to down the hierarch y (grey bars). ............. 166
17 4-10 Proximity of mothers with (black bars) and without (grey bars) their young infants (<3 months old) to troopmates during the fr uiti ng s eason. ................... 1664-11 Grooming duration (%) between female s with infants and adult troop-mates. .. 1674-12 Total duration of time (min) that mot hers with (black bars) and without (grey bars) their infants (< 3 months ol d) groomed and receiv ed grooming from troop-mates during the fruiting s eason. ............................................................ 1685-1 Troop-level comparison of fe male reproductive state and X baseline cortisol level SE (ng/g). .............................................................................................. 2045-2 Correlation between Tinkerbells cort isol levels (ng/g) and month of pregnancy. ........................................................................................................ 2055-3 Comparison of Janes mean cortisol le vels SE (ng/g) dur ing months of her pregnancy. ........................................................................................................ 2065-4 Cortisol levels of lactating females ( n = 4) at incremental stages of infant age. 2065-5 Rates ( SE) at which cycling (black bars), pregnant (grey bars), and lactating (white bars) females gr oomed and received grooming from adult female and male troop-ma tes. .......................................................................... 2075-6 Effect of infant age (3-month increments) on the rate ( SE) at which mothers (i.e., lactating females) groomed and re ceived grooming from adult females and male s. ........................................................................................................ 208
18 LIST OF ABBREVIATIONS AA Aerial alarm ACTH Adrenocorticotropic hormone BGC Between-group competition CRH Corticotropin-releasing hormone F Cortisol FPS Proximity score between a focal female and adult females without infants FWI Female with infant FwiPS Proximity score between a foca l female and females with infants GC Glucocorticoid h Hour(s) HPA axis Hypothalamic-pituitary-adrenal axis m Meter(s) min Minute(s) MPS Proximity score between a focal female and adult males ng/g Nanograms per gram RV Raleighvallen, Suri name (the study site) SE Standard error SPE Solid-phase extraction TFPS Proximity score between a focal fe male and all adult females (i.e., FPS + FwiPS) TPA Terrestrial predator alarm TPS Proximity score between a focal female and all troop-mates (adult males, adult females and immatures) WGC Within-group competition
19 Abstract of Dissertation Pr esented to the Graduate School of the University of Florida in Partial Fulf illment of the Requirements for t he Degree of Doctor of Philosophy STRESS AND AFFILIATION AMONG WILD FEMALE PRIMATES: EFFECTS OF GROUP SIZE, RISK, AND REPRODUCTIVE CONDITION IN A DYNAMIC FOREST COMMUNITY By Erin E. Ehmke May 2010 Chair: Sue Boinski Major: Anthropology My research examines the dynamics among social, ecological and reproductive challenges, affiliative behavior, and stress (as m easured by fecal cortisol) of wild female brown capuchin monkeys in Raleighvallen, Suriname. In order to better understand if and how social affiliation is used to manage physiological stress, I first identify the socioecological variables that act as primary stressors to females in a high-risk, lowenergy environment. Results indicate an adapt ive function of differential group size, particularly within an unpredictabl e ecological landscape wher e the cost/benefit regime can shift quickly. The study period proved to be a beneficial time for the smaller group observed in that the larger group females experienced increased competition, reduced allogrooming, and higher cortisol levels. Perceived pr edation risk also varied according to group size and acted as a primary stressor to females. I also focus on more fine-scaled dyadic so cial relationships and test the long-held assumption that capuchins are categorica lly a female-bonded genus. My results show that female bonding is not a stat ic, species-specific characteristic, but rather a labile strategy reflecting ecological and reproductive parameters. Relationships, measured
20 through grooming and proximity patterns, betw een females were transient and occurred primarily in response to infant presence. Females were arguably more strongly bonded to adult males, and the alpha male in particu lar, possibly in an effort to acquire increased predator protection and tolerance at food resources. Finally, I investigate if female cortisol levels co-vary with shifts in their social relationshi ps. Intriguing patterns emerge implicating the potential stress-reducing effects of strengthened relationships, particularly with male conspecifics. My study explores (1) the interaction of social and ecological parameters that affect the daily lives and dec ision-making processes of wild primates, contributing to the understanding of emergent variability in so cial behavior among group-living animals, and (2) the association between stress and soci al behavior. Since chronic stress is detrimental to health and fitness, behaviora l responses aimed at reducing stress, namely social bonding and investment, ar e adaptive and provide a mechanistic understanding of the evol ution of sociality.
21 CHAPTER 1 INTRODUCTION AND LI TERAT URE REVIEW General Research Goal and Specific Objectives My research asks, what are the b enefits of affiliation? As there are opportunity costs to social behavior, in what social, reproductive, and ecological circumstances do individuals strategically use social relationshi ps? Specifically, I investigate if and how female primates use non-random affiliative behav ior (i.e., social bonds) as a means to mediate or circumvent physiological stress. The concept of social bonds is frequently (a nd abstractly) referred to, but the topic is rarely a direct focus of empirical study. Additionally, the term social bond is an ambiguous one, further complicating its understanding and undermining its comparative value within biological systems. The goal of my dissertation is three-fold: (1) to provide an operational definit ion of social bond, one that can be replicated and studied in other systems; (2) to present data of comparative use; and (3) to test hypotheses regarding the proximate mechanisms and subsequent effects of female relationships. To accomplish these goals, I use brown capuchin monkeys ( Cebus apella ) as my study system. Cebus is arguably the best-studied ge nus of New World non-human primates. The social, ecologic al, and reproductive lives of wild C. apella, however, became largely over-generalized as a result of research based at only a few field sites. My study site represents a little-studied ec osystem for brown capuchins, and as such, preliminary data indicate striking behavioral deviations from the expected norm. Further, there are no published reports regarding the endocri nological basis of (nonsexual) behavior within wild female C. apella.
22 In the following chapters, I examine the re lationship between sociality, affiliation, and stress. In this introductory chapter, I review the physiology of stress, synthesize theory regarding primate socioecology, and discuss the prevalenc e and evolution of mammalian social bonds as well as the bi ological basis to and social functions of grooming behavior. I then in troduce capuchin monkeys and briefly review what is known regarding the behavioral endocrinology of and affiliation among Cebus females. Literature Review The Physiology of Stress and Its Measurement Stress is a complex physiological mec hanism that integrates physiological and behavioral responses to any threat to homeostasis, whether real or perceived (Tilbrook and Clarke 2006). Over the short term, acut e stress response is an adaptive condition whereupon energy is diverted from long-term processes such as digestion, growth, immune function, and reproduction and mobilized to meet short-term needs, thus increasing an individuals likelihood of surv iving and re-establishing homeostasis. Over extended periods, however, t he metabolic costs of su stained stress rise steeply (Sapolsky et al. 2000), adversely affecting fema le health and reproductive ability, fetal development and the development of offspring social skills, and stress-coping abilities (Sachser et al. 1998; Bardi et al. 2005; Sapolsky 2005; Shively et al. 2005; Wadhwa 2005). In humans, stress has been linked to numerous adverse conditions, including depression, anxiety, cancer, cardiovascu lar disease, diabetes and dementia (Charmandari et al. 2005). When an individual is confr onted with a stressor, be it in trinsic or extrinsic, the hypothalamic-pituitary-adrenal (HPA) axis is activated. Within seconds, corticotropin releasing hormone (CRH) is produced in the hypothalamus and acts on
23 the central nervous system to signal acut e stress; CRH then stimulates the anterior pituitary gland to produce adrenocorticotropi c hormone (ACTH), which in turn signals the adrenal cortex to secrete the glucocorticoids (GCs) ma inly cortisol (in primates, carnivores, and ungulates) or corticosterone (i n rodents, birds, and reptiles) (Touma and Palme 2005). These GCs are steroid hormones that coordinate an individuals physiological and behavioral response to the st ressor, acting to divert energy from long term processes (e.g., digestion, growth, immune func tion and reproduction) into the mobilization of energy necessary to m eet the demands of the challenge (Sapolsky et al. 2000). Glucocorticoids then provide a negative feedback mechanism by regulating the secretion of CRH and ACTH, t hus down-regulating their own production (Tilbrook and Clarke 2006). Steroid hormones are released as they are produced into the bloodstream where they then act on their target tissues. While blood plasma concentrations of GCs may be used as a measure or indicator of stress, such sampling is associated with numerous potential problems. Because of the pulsatile nature of GC secr etion, plasma GC concentration can vary significantly withi n a matter of minutes and measure only a single point in time. GC production may al so show diurnal rhythm patterns. These factors introduce dramatic interand intraindividual variation in plasma hormone concentrations at any given time and make t he data difficult to interpret, especially on the individual level. Additi onally, the restraint techniques necessary to obtain a blood sample are stressful on the animal, and thus the sampling methodology itself increases the immediate concentration of plas ma GCs (Palme et al. 2005).
24 Fecal sampling is a feasible alternative fo r measuring GC concentration that offers significant advantages over blood sampling, with comparable results (Boinski et al. 1999b). A non-invasive technique, fecal sampling can be done with minimal disturbance to the animal, and the feces c an be collected frequently and with (relative) ease. Further, circulating steroid hormones are metabolized extensively by the liver and excreted as conjugates (Palme et al. 2005), a process that effectively pools hormone levels over several hours and minimizes sh ort-term hormonal fluctuations and rhythmic patterns (Goyman et al. 1999). Therefore, fe cal corticosteroid metabolites provide a useful indicator of genuine and prolonged physiol ogical stress response s (Boinski et al. 1999b). This process of metabolization and excretion necessitates a species-specific time lag which is related to the animals in testinal transit time from duodenum to rectum (Touma and Palme 2005); in non-human primates, this lag time is typically less than 48 hours (Whitten et al. 1998): 22 hours in macaques ( Macaca fascicularis ), 24 hours in common marmosets ( Callithrix jacchus ), 26 hours in chimpanzees ( Pan troglodytes ) (Bahr et al. 2000), and 36 hours in yellow baboons ( Papio cynocephalus ) (Wasser et al. 2000). Despite the acknowledged relationship between stress hormones and genetic, environmental and social factors (Touma and Pa lme 2005), variation is often controlled for in a laboratory environment. Thus, our understanding of stress and its relevancy for an animal in its natural environment is lim ited (Reeder and Kramer 2005). I now turn my focus to a review of the evolutionar y relationship between an animals natural environment and its social behavior.
25 Primate Sociality, Competition, and Soci al Structure: A Historic Perspective Acknowledgement of the insepar able link between an organisms ecological and social environment became widespread in the 1970s when empirical studies demonstrated that social struct ure varies with local ecologic al conditions (Eisenberg et al. 1972; Clutton-Brock and Harvey 1977; Lott 1984). At that time, the role of the male primate was emphasized, and female social behavior was largely overlooked. Since then, however, females have taken center st age, as it is generally accepted that the social and sexual strategies of females determine the behavioral strategies of males (Emlen and Oring 1977). Wrangham (1980) developed the first ecological model of primate social behavior, emphasizing between-group competition (BGC) as the primary mechanism underlying the evolution of primate social ity. Strong BGC favors indivi duals that develop a social bond, facilitating cooperative displacement of rival groups from resources; this encouraged the development of philopatry, with one sex remaining in its natal group to maintain such cooperation through kinship a nd familiarity. As such, a female-bonded species was defined as one in which females form cohesive matrilineal groups characterized by affiliative interactions, su ch as proximity, grooming, and the formation of coalitions. Van Schaik (1989) revised Wranghams model of social evolution, arguing that predation is the mechanism by which socialit y evolved. An increase in the number of collective individuals progressively incr eases the number of eyes scanning the environment for predators. T herefore, with an increase in group size, an individuals likelihood of survival improves, and anti-predatory vigilance e ffort required per individual decreases (Elgar 1989; Quenette 1990). Subsequent to the formation of groups,
26 within group competition (WGC) became the determining factor in the development of social relationships, with females alleviati ng the costs of WGC thro ugh the formation of a linear dominance hierarchy and coalitionary support. Furt her, the term bond was uncoupled from philopatry and became a more general condition characterized by affiliation and the formation of agonistic alliances. Mammalian Social Bonds: A Review Fine-scaled analyses of the contextual and functional basis for mammalian social behavior are limited, largely, to pri mates (K utsuskake 2009). Possibly because of this, female-bonded social systems are known to occur only in isolated and rather exceptional cases among non-pr imates (Shultz and Dunbar 2007, p. 2435). Further, little attempt has been made to define behavioral measures of social bonds in non primates (Shultz and Dunbar 2007); for most non-primate taxa, however, the primate measures of allogroom ing and proximity are cited and used as indices to signify bondedness. Non-primate mammalian taxa in which differ entiated, affiliative intrasexual adult relationships have been the focus of empiri cal study is largely limited to hyenas ( Crocuta crocuta), elephants ( Loxodonta africana ), and horses ( Equus spp.). Among hyenas, females associate more with same-sex ed partners than do males (Smith et al. 2007). High-ranking females are more gregarious than low-ranking females (Holekamp et al. 1997; Smith et al. 2007), and subordinate females initiate associations more often than do dominant females (Smi th et al. 2007). Further, Crocuta demonstrates potentially costly greeting ceremonies that are thought to serve a socially cohesive function, likened to that of grooming behav ior among primates (Drea and Frank 2003), and dyadic greeting behavior follo ws the same patterns of association (East et al.
27 1993). In response, dominant females better to lerate those subordinates that initiate social interaction, providing them wit h more dependable access to food and reduced rates of aggression (Smith et al. 2007). Smith et al. (2 007) suggested that female Crocuta crocuta minimize the costs of group living by directing affiliative social behavior towards outranking conspecifics that provid e return benefits. Female elephants demonstrate stable, non-random patterns of soci al behavior, particularly within family units. Social relationships of non-kin, how ever, show temporal shifts, especially in response to seasonal changes in resource ava ilability (Wittemyer et al. 2005). Among feral horses, social relationships betw een females involve preferred partners who allogroom and maintain proximity (Heitor et al. 2006; Cameron et al. 2009), and social bonds between females may be more important to the social stability of the group than are intersexual pair bonds (Curry et al. 2007) Cameron et al. (2 009) found that, among unrelated females in a New Zealand population, more gregarious mares received less harassment from groupmates, which may explain the positive relationship between female social integration and reproductive success. These findings contrast with those of Heitor et al. (2006), w ho demonstrated that among feral horses in Portugal, mares with stronger affiliative relationships we re not less aggressive towards one another, regardless of kinship. Despite the presence of differentiated, dyadic social relationships in female hyenas, elephants, and horses, there exist fundamental differences in the social structures between these taxa and most primate taxa. Hyenas and elephants both have fission-fusion societies, not stabl e social groups that charac terize the females of most primate species. The dynamics of constantly shifting sub-groups provide more flexibility
28 in partner choice decisions compared to i ndividuals residing in more stable, cohesive social groups (Smith et al. 2007). In group living species with relatively stable compositions, the history of repeated social interaction between group members shape each dyadic (and polyadic) relationship, pro ducing distinct selective pressures which cannot be observed in species with unstabl e group membership (Kutsukake 2009). Spider monkeys ( Ateles spp.), one of the only primates to demonstrate a fission-fusion society, provide an ideal example as to the effects of such an unstable social system on individual social relationships. Female Ateles show a high level of association, but little selectivity in their association partners. In fact, females associate with one another no more than randomly expected. Instead of diff erentiated relationships, females passively aggregate in subgroups in response to ecological variables, such as the size and prevalence of particular food patches (Ramos-Fernandez et al. 2009). Horses, by contrast, reside in stable so cial groups; both sexes disperse, however, and so females are not expected to develop and maintain the life-long female associations such as those observed in female-philopatric primates. This likely has profound effects on the mechanisms perpetuating social relationships. In squirrel monkeys ( Saimiri spp.), for example, inter-specific va riation in female dispersal patterns affects the strength and st ability of female social relationships. In S. boliviensis females are philopatric and form stable, lin ear dominance hierarchies and long-term social bonds. In S. sciureus female dispersal is flexible (females may remain in natal group or emigrate), and as such, social beh avior is rather erratic. Although females form stable, linear dominance hierarchies, severe aggression occurs more frequently and female-female coalitions are less common in S. sciureus than in S. boliviensis
29 Finally, S. oerstedii is characterized by female disper sal. In response, females do not develop a dominance hierarchy, participate in coalitions, or form long-term social bonds (Boinski et al. 2002). In order to better understand the focus on primates in regards to social bonds, I next review the evolut ionary and biological mechanisms underlying female bondedness. Mammalian Social Bonds: An Evolutionary Perspective The evolutionary basis to social bonds is the mother-infant bond. The ontogeny of mammalian infants, both preand post-natally requires extensive specialization on the part of the mother, e.g., plac entation, lactation and intensive maternal care. To ensure successful nurturing of offspring, the mate rnal brain releases oxytocin, a neuropeptide hormone that facilitates milk let-down and olfa ctory recognition of offspring, stimulating the process of maternal bonding (Curley and Keverne 2005). The development of other social relati onships seems to be regulated by the pre-existing bonding mechanism (Curle y and Keverne 2005). In small-brained monogamous mammals such as prairie voles ( Microtus ochrogaster ), the release of oxytocin also occurs post-mating, enabling t he female to form an exclusive olfactory partner preference (i.e., an endur ing pair bond) (Getz and Cart er 1996; Cho et al. 1999; Sanchez-Andrade and Kendrick 2009). Thus, the biological mechanisms underlying bond formation in small-brained mammals ar e hormone-dependent and rely heavily on olfactory information; as such, their most significant social relationships are between mother and infant and between mating partner s. Haplorhine primates (the monkeys and apes), however, diverged from the ances tral mammalian pattern of olfactory reliance and hormonally-facilitated social relationships (Curley and Keverne 2005).
30 The shift from a reliance on olfactory to visual cues is an important difference between haplorhine and other mammalian social bonding, and the higher cognitive functioning of primates enables complex social relationships to develop outside of hormonal priming. Even sexual behavior and parental care are less controlled by hormones in primates, as sexual interact ions frequently occur in a non-reproductive context and post-partum care extends well beyond the period of lactation dependency. Such a release from hormonal determinant s both enables alloparenting behavior and requires social learning of parental and so cial skills (Curley and Keverne 2005), and also expands the breadth of important adult relationships beyond mating partners (Dunbar 2009). As the formation of social bonds am ong primates is more cognitive than physiological, primates must continuously rein force their social bonds through affiliative group interactions. In order to sustain such a group dynamic, a physiological reward mechanism is required. Just as tactile st imulation (parturition and copulation) in smallbrained mammals stimulates the release of oxytocin (Curley and Keverne 2005), the tactile stimulation of primat e grooming initiates the rel ease of the opioid neuropeptide, -endorphin (Keverne et al. 1989). This mechani sm of rewarding social interactions promotes the development of individual relationships (i.e., social bonds) among primates (Curley and Keverne 2005). The Biological Basis of Grooming All primates groom (S humaker and Beck 2003) This generalized characteristic of the order Primates, however, applies to hygienic (self) groomi ng (i.e., removal of ectoparasites, dirt) and not necessarily to groom ing that provides a social function. For most primates, however, it is generally accept ed that grooming is social (Schino 2001).
31 Social grooming provides biological and physiol ogical benefits to the recipient in terms of ectoparasite removal (Zamma 2002), bet a-endorphin release (Keverne et al. 1989), reduced stress (Schino et al. 1988; Engh et al. 2006a; Crockford et al. 2008; Wittig et al. 2008) and lower heart rate (Boccia et al. 1989; Aureli et al. 1999). Social grooming may also mediate reconciliation following an agonis tic event (Leca et al. 2002) or increase tolerance around food resources (Barrett et al. 2002). Allogrooming is the most commonly observe d form of affiliation among primates (Di Bitetti 1997; Henzi and Barrett 1999). There are a few prim ates that do not commonly groom (aside from self-grooming and mother-infant interactions), such as orangutans ( Pongo spp.), owl monkeys ( Aotus spp.), squirrel monkeys ( Saimiri spp.), muriquis ( Brachyteles arachnoides hypoxanthus), and to a lesser extent spider monkeys ( Ateles spp.). Orangutan females do not allogr oom, although they may tolerate, rest, and maintain proximity with one another (T obach and Porto 2006; V ogel 2008); their lack of physical affiliation may stem from their prim arily solitary social structure. Owl monkeys, the only nocturnal anthropoid, rarely social groom. Moynihan (1964) described owl monkeys as grooming only before copulation, but Wright (1994) never observed grooming among wild owl monkeys, even before copulation. Owl monkey social behavior appears restricted by dar kness, as play and aggression are most frequent during monthly lunar cycles when the m oon is at its brightest (Wright 1994). Squirrel monkeys also rarely, if ever, engage in social grooming (Boinski 1994; Boinski et al. 2002), although they do engage in affiliative behavior known as huddling (Saltzman et al. 1991). Muriquis (a.k.a. w oolly monkeys) do not groom (Strier 1992a), and the rate of grooming among spider monkeys is low, both in captivity and in the wild
32 (Ahumada 1992; Schaffner and Aureli 2005). This may be a result of their vestigial thumb, as their hand morpholog y does not provide the dexteri ty that primate grooming typically requires (Strier 1992a). Instead, spider monkeys and muriquis demonstrate affiliation by embracing, which is suggested to fulfill the same social functions (e.g., provide access to infants, increase toler ance, and reduce the risk of aggression) that grooming provides for other primate species (Strier 1992a,b; Strier et al. 2000, 2002; Schaffner and Aureli 2005; Slater et al. 2007). Grooming is not unique to the primates. Hygienic grooming is common throughout the mammalian world and is observed frequent ly among rodents (Hawlena et al. 2008), felids (Eckstein and Hart 2000), equids (Rubenstein 1981), bov ids (Mooring and Samuel 1998a; Hart and Pryor 2004; Kohari et al. 2009), and cervids (Hart and Hart 1992; Mooring et al. 1996; Moori ng and Samuel 1998b). Alt hough not as common, a few non primate animal species demonstrate a social function to grooming as well. For example, among meerkats (Suricata suricatta ), dominant females groom the dominant males more than subordinates, and thus grooming likely functions to maintain the sexual relationship between dominant (breeding) individuals. In te rms of female-female grooming, subordinate meerkats focus on dominant females, and the duration of grooming is positively associated with frequency of aggression received from the dominant; therefore, intrasexual groomi ng among female meerkats may serve to placate the dominant individuals (Kut sukake and Clutton-Brock 2006; Madden and Clutton-Brock 2009). Grooming has also been shown to reduce social tension in horses ( Equus caballus ) (Feh and de Mazieres 1993). Among vampire bats ( Desmodus rotundus ), rates of social grooming correlate wit h rates of regurgitation (food sharing),
33 and thus is it likely that social grooming facilitates food-sharing either through honest indicators of body weight (i.e., identifyi ng which individuals are in need of food and which are able to provide it) (Wilkinson 1986) or through the exchange of services such as ectoparasite removal for food sharing. The Social Functions of Grooming Grooming may be used strategically as a commodity to be traded, as suggested above by meerkats, horses, and vampire bat s. The grooming-for-aid hypothesis suggests that the benefits of grooming are exchanged directly fo r coalitionary support during interand intra-group competition. Since dominant individuals are expected to be more effective allies and therefore more attrac tive social partners, it is predicted that grooming should be directed up the hierarchy, with subordinate females competing to groom higher-ranking females (Seyfarth 1977). Empirical evidence supporting this hypothesis, however, is inconsistent and focused more on Old World monkeys (Sade 1972; Seyfarth 1977; Stammbach 1978; Silk 1982; Seyfarth and Cheney 1984; Schino 2001). Conversely, several studies report the opposite trend, with grooming directed down the hierarchy, possibly as a means fo r the dominant female to secure support from subordinate females (O'Brien 1993; Linn et al. 1995; Di Bitetti 1997; Parr et al. 1997; Franz 1998; Henzi et al. 2003). Other studies found no link between dominance and grooming directionality (de Waal and Luttr ell 1986; Thierry et al. 1990; Bernstein and Cooper 1999; Matheson and Bernstein 2 000; Schino et al. 2009). Further, evidence supporting the direct relations hip between grooming investment and coalitionary aid is minimal at best and o ften deemed circumstantia l (Henzi and Barrett 1999; Silk et al. 2004; but see Schino 2007).
34 Grooming may also be exchanged for groomi ng itself (i.e., reciprocal grooming) (Silk et al. 1999; Payne et al. 2003; Schino and Aureli 2008; Fedurek and Dunbar 2009), as a means to gain access to infants (O 'Brien 1993; Muroyama 1994; Perry 1996; Di Bitetti 1997; Manson et al. 1999; Henzi an d Barrett 2002), to pr omote social and food related tolerance (Kapsalis and Berman 1996; Ventura et al. 2006), and to establish, maintain, or repair social relationships (Stammbach and Kummer 1982; Dunbar 1996; Lehmann et al. 2007; Fedurek and Dunbar 2009). The value of grooming as a commodity or currency varies not only by services rendered, but also by the individuals involv ed and the current socioecological milieu of the biological marketplace. In order to maximize ones return, an individual would be expected to direct affiliation towards those partners that potentially provide the most benefits. The benefits exchan ged, however, need not be the same (i.e., animals may trade in different currencies), as individuals may vary in t heir ability to provide benefits of one type of currency over another. Additionally, the value of different commodities varies depending on ecological conditions (Schino and Aureli 2008). For exam ple, when the dominance hierarchy of a group is weakly expressed or enforced, gr ooming (a rank-independent currency) has greater value and is typically exchanged fo r grooming (Henzi et al. 2003; Schino and Aureli 2008). Under condi tions of stronger WGC, however, dominance hierarchies become steeper and grooming recipr ocation decreases (Barrett et al. 2002); instead, grooming becomes increasingly directed up t he hierarchy (Ram et al. 2003) as other currencies, such as tolerance, become mo re valuable as exchan ge commodities for grooming. In a meta-analysi s of grooming distribution among various primate taxa,
35 Schino and Aureli (2008) found a flexible trade-off between the exchange of grooming for rank-related benefits and the need to trade grooming for its more direct biological benefits. Henzi and Barrett (1999) suggested that primate allogrooming provides great social significance based on evidence that the behavior was conserved in the presence of other demands. Dunbar (1991) determined that the social func tions of grooming, however, primarily pertain to Old World primates based on data showing that their grooming rates correlated with group size, whereas grooming rates among New World monkeys were more associated (but not signi ficantly) with body size. As the grooming behavior of New World monkeys has become more thoroughly studied, however, the demonstrated social functions of platyrrhi ne grooming are indeed com parable to that of catarrhines (OBrien 1993; Perry 1996; Di Bitetti 1997; Schino 2007; Schino and Aureli 2008). In particular, the capuchin monkey ( Cebus spp.) provides a study system unique to platyrrhines and more akin to catarrh ine monkeys, thus bridging the gap between New and Old World primates (Parr et al. 1997). Cebus: Ecologically Diverse, Socially Similar? Capuchin monkeys are one of the most widely distributed genera of New World primates, found throughout most of Central and South America. They inhabit a diverse range of habitats including tropical rainfores t, swamp forest, and deciduous forest with seasonal, desert-like conditions (Fragaszy et al. 2004). Traditionally, the genus is composed of four species: C. albifrons, C. apella, C. capucinus and C. olivaceus (Hill 1960). In recent years, howev er, the number of recognized s pecies ranges from five to eleven, with much of the controversy focusing on species versus subspecies distinctions of C. apella (Fragaszy et al. 2004). For consistency with the majority of
36 published literature and to side-step the ta xonomic confusion until much of the controversy is settled, I refer to the tr aditionally recognized species (Table 1-1). Cebus spp. are best described as opportunist ic and adaptable; in fact, their habitat-use patterns are so flex ible that they defy specie s-specific characterization (Fragaszy et al. 2004). Adult social relati onships, however, are typically described through assumed species norms. For example, all Cebus species are classified as strongly female-bonded (Janson and Boinsk i 1992; Fedigan 1993; O'Brien 1993; Perry 1996; Di Bitetti 1997), perhaps mostly because capuchins are generally femalephilopatric but also because of strong fema le-female grooming patterns observed at a few field sites. In a genus noted for its ability to adapt to diverse ecological conditions, should capuchin social behavior be generalized at the species level? The short answer: No. Recent findings from Brazil introduced a brown capuchin population that lacked female grooming bonds (Izar 2004), and prelim inary observations of female C. apella in Suriname suggested both temporal and individual variation in intrasexual relationships (Boinski, unpublished data). Further, despite a generalized tendency for fe male capuchins to prefer female grooming partners, there exists no interor intra-specific consensus on the value of grooming as a currency or how it is used strat egically in the biological marketplace. The females in some groups groom up the hier archy (Perry 1996), while in other groups females groom down the hierarchy (O'Brien 1993; Parr et al. 1997), and some females demonstrate no relation between grooming part ner preference and rank (Manson et al. 1999; Schino et al. 2009). There also exists no clear relation between grooming and
37 coalitionary support (Perry 1996; Di Bitetti 1997; Manson et al. 1999; Schino et al.,2009), especially after controlling for r ank and kinship. Schino et al. (2009) suggested that Cebus grooming behavior is so heterogeneous because of differences in group size and the varying number of matrilines between groups. They proposed that grooming could be expected to be exchanged for agonistic support only when the group is comprised of more than two matrilines (i.e., when mean relatedness is lower). Otherwise, in smaller groups, there is no opportunity for t he exchange of grooming for support since females preferentially support cl ose kin over nonkin. Therefore, while grooming may serve as a means to ensure agonistic support for some Cebus females, empirical data more consistently support the notion that grooming is a commodity to be traded for the benefits of groomin g itself (Schino and Aureli 2008) and to gain access to young infants (O'Brien 1993; Perry 1996; Di Bitetti 1997; Manson et al. 1999). The hormonal correlates of social behavior for wild brown capuchin females remain unknown. Published reports of Cebus behavioral endocrinology are scant and limited to C. apella males (Lynch et al. 2002), captive C. apella females (Linn et al. 1995; Carosi et al. 1999; Carosi and Visa lberghi 2002; Lahoz et al. 2007), and wild C. capucinus females (Carnegie et al. 2005). In fa ct, female behavioral endocrinology has been studied only in terms of sexual behavior and reproductive physiology (Linn et al. 1995; Carosi et al. 1999; Carosi and Visal berghi 2002; Carnegie et al. 2005; Lahoz et al. 2007). To date, no repor t has been published on the dynamics of stress in terms of social and ecological variation for wild female C. apella Dissertation Overview : What Lies Ahead In the following four chapter s, I empirically examine groupand individual-level hypotheses regarding sociality, affiliation, and physiological stress (Table 1-2). In
38 Chapter 2, I explore the cost s and benefits of sociality in te rms of differential group size, specifically addressing how predation risk, a seasonal env ironment and social behavior vary with group size. In Chapter 3, I examine t he intrinsic effects (i.e., cortisol levels) of group size and test group-level hypotheses regarding the relation ship between group size, female cortisol levels, predation risk, seasonality, and intragroup agonism. In Chapter 4, I study fine-scaled social rela tionships on the individual level and test hypotheses regarding the consistency of dyadic relationships between females (i.e., are the females truly bonded?). I also explore t he relation between female social behavior, ecological pressures, and r eproductive state. Finally, in Chapter 5 I examine the hormonal mechanisms driving sociality. Do i ndividual females moderate incurred social and ecological stress by strategically adjusting their social relationships?
39 Table 1-1. Socioecological comparison of Cebus populations at various field sites. C. apella (Nouragues, French Guiana) C. apella (Brazil) C. apella (Iguazu, Argentina) C. apella (Manu, Peru) C. apella (Raleighvallen, Suriname) C. capucinus (Lomas Barbudal, Costa Rica)C. capucinus (Santa Rosa, Costa Rica)C. olivaceus (Venezuela) C. albifrons (Colombia) Site description Seasonal primary rainforest, mixedhabitat Atlantic forest Seasonal subtropical forest with bamboo patches Seasonal lowland flood-plain Highly seasonal & variable, mixed-habitat Seasonal tropical dry deciduous forest Seasonal tropical dry forest; limited water Seasonal tropical dry deciduous forest Gallery forest Predation risk n.a. Low Low High High n.a. n.a. n.a. n.a. Resource competition n.a. High WGC and BGC High WGC and BGC High WGC, low BGC High BGC; variable WGC High WGC n.a. n.a. n.a. dominance hierarchy n.a. tolerant linear n.a. weakly linear linear; stable n.a. stable; matrilineal n.a. -bonded (grooming patterns) n.a. no yes n.a. no yes yes yes n.a. Hierarchal grooming direction n.a. n.a. down n.a. no trend up no trend down n.a. grooming competition n.a. no no n.a. no no no no n.a. Published literature Zhang 1995a,b; Zhang and Wang 1995, 2000 Izar 2004 Janson 1996; Di Bitetti 1997, 2001; Hirsch 2002 Janson 1985, 1988a,b, 1990a,b; van Schaik and van Noordwijk 1989 current study Perry 1996; Manson et al. 1999; Vogel and Janson 2007; Vogel et al. 2007 Perry et al. 2008 Fedigan 1993; Manson et al. 1999 O'Brien 1991, 1993 Defler 1982 Notes: WGC: within group competition; BGC: be tween group competition; n.a.: data not available.
40 Table 1-2. Overview of di ssertation chapters 2 to 5. Chapter Hypotheses 2. The effect of group size on the social and ecological milieu of female brown capuchins. 1. Daily travel distance increases with group size. 2. Individual foraging time in a preferred, monopolizable food tree decreases with increased group size. 3. Predator alarm frequency (per individual) is inversely associated with group size. 4. Rates of female agonism increase with group size; group size affects the strength of female dominance hierarchies. 5. Rates of female grooming increase with group size. 6. Female brown capuchins in a large troop demonstrate more seasonal variation in patterns of grooming and agonism than do the females in a small troop. 3. Group size and stress: The effects of a high risk, low energy environment on female brown capuchins. 1. Group size affects the level of stress females incur due to predation risk. 2. Group size affects the levels of stress females incur during food shortages. 3. Rank-related stress is a factor of group size. 4. Female-bonded primates? Grooming and proximity patterns of female brown capuchins in Suriname. 1. Social bonds between adult female brown capuchins are variable over time. 2. Dyadic female relationships vary in response to ecological pressures. 3. Relationships among females strengthen with the presence of an infant. 5. Do female brown capuchin monkeys use affiliative behavior to mediate stress? 1. Female cortisol levels are inversely associated with rates of affiliative behavior. 2. Female cortisol levels vary with monthly variations in affiliation. 3. Female cortisol levels vary with reproductive condition. 4. Female reproductive condition affects affiliative behavior.
41 CHAPTER 2 THE EFFECT OF GROUP SIZE ON THE SOCIAL AND ECOLOGICAL MIL IEU OF FEMALE BROWN CAPUCHINS Introduction There is consensus that the benefits of so ciality are accompanied by the costs of group living. Numerous studies of social animals investigate the various costs (e.g., increased intragroup competition, more cons picuous to predators) and benefit s (e.g., improved intergroup competition, foragi ng benefits, better predator detection and defense) associated with group living, typically by comparing differences in behavior and reproductive success among groups of differ ent size (van Schaik 1983; van Schaik and van Hooff 1983; Pulliam and Caraco 1984; Watts 1985; Cheney and Seyfarth 1987; Dunbar 1987; Janson 1988b; Robinson 1988; Isbell and Young 1993; Takahata et al. 1998; van Noordwijk and van Schaik 1999; Chapman and Chapman 2000; Krause and Ruxton 2002; Boinski et al. 2003; Keeling et al. 2003; Pri de 2005a; Isvaran 2007; Silk 2007a; van Belle and Estrada 2008). The expec tation is that individuals maintain membership in groups of favorable size to maximize fitness (Wrangham 1980); individuals resident to groups t hat are smaller or larger than optimal size, therefore, will reproduce less successfully than those living in intermediate-sized groups (Silk 2007a). What are the short-term ecological and social variables associated with group size that result in such a long-term fitness differential? This question is especially important for the females of female-philopatric species, as they tend to remain in their natal troop regardless of its size. In this chapter, I test hypotheses r egarding the costs and benef its of disparate group sizes using a familiar female-philopatric study species resident to one of the least studied biogeographic regions of the Neotropics. The brown capuchin monkey ( Cebus
42 apella ) is one of the most geographically divers e New World primates, and the species is well-known from populations studied in Peru and Argentina. Do the established brown capuchin norms hold true in an ecosystem characterized by a stringent and highly variable cost/benefit regime? Indeed, the two study groups on which my research is based are not of optimal size (i.e., intermedi ate (Silk 2007a)), but represent opposite ends of the continuum for group size (Table 2-1). Hypotheses Daily Travel Distance In creases w ith Group Size The ecological constraints model, which pred icts group size as a function of travel costs (Snaith and Chapman 2007), is based on the assumption that an increase in group size results in an increase of within-group resource competition (WGC). In response, the group may expand the area that must be searched in order for group members to acquire adequate nutritional intake (Eisenberg et al. 1972; Bradbury and Vehrencamp 1976; van Schaik et al. 1983a; Janson 1988a; Janson and Goldsmith 1995; Chapman and Chapman 2000; Ganas and Ro bbins 2005) until a point is reached at which travel costs exceed energetic retu rns from the environment, and smaller group size becomes advantageous (Chapman and Chapman 2000). Specifically for brown capuchins, Janson (1988a) found that of small groups, foraging efficiency (food ingested per unit distance traveled) was up to four times greater than that observed of large groups. Individual Foraging Time in a Prefe rred, Monopolizable Food Tree Decr eases with Increased Group Size Food patch size influences the degree to wh ich individual foraging efficiency is affected by group size, as small patc hes are more quickly depleted with more
43 individuals feeding. Janson ( 1988a) found that, in large food patches, individual feeding time was not affected by group size, whereas in small patches, individual feeding time decreased with group size. Predator Alarm Frequency (per individua l) is Inversel y Associated with Group Size Perceived predation risk affects an i ndividuals foraging and social behavior (Chapman and Chapman 2000). Because an increase in group size increases the total amount of vigilance and likelihood of predator detection, i ndividuals in larger groups are able to reduce the amount of time spent vi gilant (Isbell and Y oung 1993; Isvaran 2007; Rieucau and Martin 2008) and spend more time foraging (Isvaran 2007), resting, or engaged in social behavior. The reverse may re sult, however, as small, cryptic groups may avoid predation more effectively than large, conspicuous groups (Terborgh 1983), and increased travel time and foraging effort characteristic of larger groups results in more occasions for exposure to predators (Janson and Boinski 1992). Rates of Female Agonism Increase with Group Size and Group Size Affec ts the Strength of Female Dominance Hierarchies As intragroup competition is expected to increase with group size, it follows that rates of aggression and displacement increas e with group size (van Schaik et al. 1983a) as has been empirically demonstrated (Kondo et al. 1989; Keeling et al. 2003). The presence of a dominance hierarchy, however, is widely believed to reduce intra-group aggression (Hemelrijk 1999), as individuals know their place and aggression is ritually, rather than physically, expr essed. Among farm animals, for example, aggressive behavior in large groups diminished with the presence of established dominance relationships (Rodenburg and Koene 2007). Perry et al. (2008) showed, however, that among white-faced capuchins ( C. capucinus ), the social behavior of females in small
44 groups was more affected by dominance rank than was the social behavior of females in large groups. Rates of Female Groomi ng Increase w ith Group Size Females in larger groups may either intens ify grooming in order to maintain social cohesion (Cowlishaw and Dunbar 1991), or (2) be constrained in the time available for grooming due to increased travel and foraging effort necessary to off-set increased intragroup competition within a large gr oup (Dunbar 1988). For Old World primate species, most studies agree that there is a positive correlation between time spent grooming and group size (Dunbar 1988; Cow lishaw and Dunbar 1991; Isbell and Young 1993; Lehmann et al. 2007). For New Worl d monkeys, however, the association between group size and grooming is inconsist ent. OBrien (1993) reported that, similar to catarrhines, grooming frequency among female C. olivaceus was affected by group size, with females in a large group involved in more grooming than females in a small group. He suggested that t he reduced social complexity (fewer matrilines and higher degree of relatedness) of small groups result ed in decreased competition and thereby decreased female grooming. However, Ma nson et al. (1999) found no relationship between C. capucinus group size and dyadic female grooming rates. Female Brown Capuchins in a Large Tr oop Demon strate More Seasonal Variation in Patterns of Grooming a nd Agonism than Do the Females in a Small Troop For most New World monkeys, the propor tion of time invested in grooming and other social behaviors is seasonally variabl e (Di Bitetti 1997; Freese and Oppenheimer 1981; Kinzey and Wright 1982; OBrien 1993), due mo stly to variation in food availability and the opportunity cost of increased for aging effort (Terborgh 1983; Janson 1988a). Variation in resource availability is most likely to affect groups characterized by WGC
45 (i.e., larger groups); therefor e, individuals resident to la rger groups must vary their social strategies in response to ecological variation. Methodology Site Description The Guianan region, comprised of Frenc h Guiana, Suriname and Guyana, is a little-studied region of South America. This biogeographic area is a dynamic environment with intense annual and seasonal variat ion. Additionally, the Guianan soils are nutrient-poor and highly weathered as t hey are derived from Precambrian bedrock (Norconk et al. 1996). The study site was located in Raleighval len (RV), Suriname, a 7812 km2 reserve consisting of primary tropical ra in forest in the 1.6 million ha Central Suriname Nature Reserve. The main study site measured approximately 2 km2 and was covered by an extensive trail system RVs flora and fauna are effectively undisturbed in the historic al period, and an intact a rray of potential predators and competitors are present (Rei chart 1993). Much of RV is composed of expansive bamboo patches and dense liana forest that restricts visibilit y (Boinski et al. 2003). Fleshy fruits are low in abundance and available fruit resources typically occur in small patches (<5 m diameter) (Boinski et al. 2002). One palm species, Maximiliana maripa is particularly important to t he food competition regime of C. apella in RV. From December through June, this fruit tree provides a high quality resource (rich in digestible proteins, lipids, and carbohydrates) in dense, small clusters (<1 m diameter) (Boinski 1999). A troop may feed in a single M. maripa cluster for several hours, during which intra-group competition intensifies and dom inance interactions are prevalent.
46 Study Animal The brown capuchin monkey ( Cebus apella ) is one of the four traditionally recognized species in the genus Cebus (Hill 1960). Capuchins li v e in multi-male, multifemale polygamous groups that are typically female philopatric. Adult males and females maintain separate dominance hierar chies; in general, males are dominant to females. C. apella social groups are comprised of 12 to 27 individuals, with an average troop size of 17 members. Troop membersh ip and relationships remain relatively stable, except for subadult males who emigrate from their natal group and may subsequently transfer several times. Brown capuchins reside in a variety of habitats encompassing a large geographic range from Columbia to Argentina (Fragaszy et al. 2004). C. apella are omnivores and arguably the most labile generalists of New World monkeys. The bulk of their diet consists of fruit supplemented with seeds, vegetation, arthropods, and vertebrates (Terborgh 1983). Study troops A and B were two of f our brown capuchin groups commonly observed within the study site, and the primar y study subjects were the 10 adult females resident to these two groups. For most of the study, Troop A consisted of 27 individuals, seven of which were adult fema les; Troop B consisted of nine individuals, three of which were adult females. Duri ng the study, one pregnant female disappeared from Troop A (her disappearance coincided with her predicted du e date, and thus it is possible her death was related to giving bi rth), and two females disappeared from Troop B (cause unknown) (Table 2-2). Relate dness among the adult troop members is unknown; however, genetic analysis of the brow n capuchin population in RV is currently underway. The study troops had been habituated to human observers since 1998, and all individuals were recognize d based on body size, color patterns, and other identifying
47 characteristics such as scars, moles, and ear shape. Throughout the study period, inter-observer reliability tests were conducted monthly to ensure consistent and accurate identification of all troop members. Data Collection From January to December 2006, I and/or fi eld assistants followed Troop A for up to 10 h each day, beginning at sunrise. Troo p A was followed, on average, 28 days per month, composing a total of 2284 h of observati on (not including out of view time). I followed Troop B, on average, 8 days per mo nth, composing a total of 495 h of observation (not including o ut of view time). General group charac teristics such as troop location, travel distance, foliage cover, dispersion, height in canopy, and diet we re collected via group scans at 15-min intervals. These data provided comparison of the troop-level char acteristics of the females in Troop A with the females in Troop B. Each groups total daily travel distance was calculated using a grid of 50 m2 quadrants overlaying a map of the study site. Cebus apella social behavior (agonistic, groom ing, and sexual) accounts for a small fraction, 1% (Pollard and Blumstei n 2008); <10% (Zhang 1995a) of their daily activity budget. Therefore, ad libitum sampling (Altmann 1974) provided a more efficient and complete data collection technique for capturing social interactions than did a more structured sampling regime. To r educe observation bias, I moved constantly through the groups to locate non-visible indivi duals and record their activities. Due to opportunistic sampling, as well as the reduced visibility within the RV forest, the recorded frequencies of social behavior ar e likely underestimates of the true frequencies. Because of the large number of hours over which behavioral interactions
48 were recorded, however, I am confident t hat the data provide strong indication of underlying patterns. Grooming and agonistic behaviors (Table 2-3) as well as other opportunistic occurrences such as predator interactions, were noted ad lib Only agonistic and grooming bouts involving at least one female (as actor or recipient) and where both participants could be identified (at least to the age and sex class) were included for analysis. All bouts were counted as dyads: If two monkeys were grooming the same individual concurrently, the episode was scored as two actor-recipient grooming bouts; if agonism involved a coalition agai nst a third party, only the in itial instigator and victim were scored. Agonism and grooming involving females in estrus (as actors or receivers) were not included in analysis to pr eclude confounding a females interest in social partners with her sexual intere sts (an obvious bias towards males). Data Analysis Behavioral rates Data were grouped int o monthly and seaso nal averages, as follows: fruiting season (February to April), wet season (May to July), dry season (August to October), and transition season (November to January). To control for differences in troop size when comparing behavioral rates of the females in Troop A wit h the females in Troop B, M. maripa feeding, alarm calls, grooming, and agonism were expressed as rates per individual, assuming all individuals were equal. Although such an assumption does not reflect reality since dyadic interactions and feeding rates were likely affected by intragroup dominance relationships, the assu mption was applied to both troops and simply acted as a control for differenc es in group size and composition.
49 M. maripa feeding rates (min/h) and alarm call rates (# alarms/h) were expressed at the troopand individual-level. Alarm ca lls were often only heard and could not be ascribed to a particular individual; similarly, M. maripa feeding sessions were chaotic and characterized by high turn-over rate an d discontinuous feeding by individuals. Further, the fruiting clump wa s often at least partially obscured from view making accurate identification of feeder s difficult. Therefore, I calculated rates of alarm calls and M. maripa feeding to describe the average individual based on the number of active group members. When calculating the average rate of M. maripa feeding per individual (min/h/individual), I included only the number of adults and subadults in the troop to form the denominator, as WGC was likely most inte nse and reciprocal between these individuals. The alarm vocalizations of Cebus apella, aerial alarms (AAs) and terrestrial predator alarms (TPAs), are audibl y distinct (Boinski et al. 1999). When calculating the rate of AAs and TPAs per individual (alarms/h /individual), the number of adults, subadults, and juveniles comprised the denominator. Female agonism and grooming rates, in com parison, were calculated to control for the number of females in each group, assu ming all females equally participate in agonistic and grooming interact ions. In essence, the average Troop A female was compared to the average Troop B female. Female agonistic rates were expressed as # bouts/h/female, with bouts involving fema les as actor and females as recipient calculated separately. Female groom rates were expressed as min/h/female, with bouts involving females as actor and females as recipient calculated separately. Dominance hierarchy For both troops, female ranks were det ermined by the outcomes of decided female-female agonistic episodes. All agonist ic bouts were categorized as either
50 decided or undecided. I considered an agonist ic bout as decided if (1) identities of participants were known, and (2) one indivi dual displayed only aggressive signals whereas the other only displayed submissive behavior, or (3) one individual displayed submissive signals, whereas the other individual display ed no agonistic behavior. All other disputes were considered to be undecided (Silk et al. 2006a). Dyadic agonistic bouts between females were constructed into a dominance matrix to determine the hierarchical nature of female re lationships (Ferreira et al. 2006). Statistical analysis Non-parametric analyses were used becaus e data failed to meet the assumption of normality, even after logarithmic transfo rmation. Homoscedasticity was tested using Levenes Test for Equality of Variances; if Levenes test indicated heterogeneous variances, I used square root transformations I used the Mann-Whitney U test when comparing two independent samples (Troops A and B) and the Wilcoxin sign-rank tests for comparing two related (intragroup) sa mples. When comparing more than two samples, the Kruskal-Wallis one-way ANOVA was used in conjunction with Mann-Whitney U tests as the non-parametric posthoc test. To investigate whether an observed frequency distribution fit an expect ed one, I used the Chi-square goodness of fit test. Analyses were performed using SPSS 11.0; the significance level for all tests (two-tailed) was set at 0.05. Mean values are presented with standard error values. Results Group Size, Home Range and Habitat Use The RV forest is composed of four habitat types: (1) expansive, dense bamboo patches (the bamboo patch pr imarily utiliz ed by Troops A and B is approximately 400 m by 200 m, centrally locat ed within Troop As home range and on the southern border of
51 Troop Bs home range), (2) high forest, (3) liana/low forest, and (4 ) swamp. The home range of Troop A, roughly 700 m2 in size, is dominated by high forest (54%) with liana/low forest comprising the secondary habi tat type (28%); similarly, the home range of Troop B, roughly 500 m2 in size, is mostly composed of high forest (44%) and liana/low forest (43%) (Figure 2-1). Although bamboo forests represent only 13% of Troop As home range (Figure 2 1), the group spent the most time (37% overall) in this habi tat type, significantly more often than did Troop B ( X2 = 177.88, P < 0.001) (Figure 2-2). Seasonal comparison further demonstrates their heavy use of t he bamboo patches: In every season except the fruiting season, Troop A s pent the most time in the bamboo, most notably during the dry season (Figure 2-3). Troop B, however, spent most of their time significantly more often than Troop A ( X2 = 63.86, P < 0.001) in the liana/low fore st (49%) (Figure 2-2), a pattern that holds true for ev ery season except the transiti on season. Their use of the high forest remained relatively consistent throughout the seasons while their use of the bamboo was more variable, peaking in the transition season (Figure 2 4). Canopy cover was estimated on a scale from 0 to 3 with 0.5 increments such that the score increased with vegetation density (Boi nski et al. 2003). M ean cover ( SE) for Troop A and Troop B during the study per iod was 2.41 0.03 and 2.34 0.02, respectively. Seasonal cover was similar between the troops, with the largest differential of 0.20 during t he dry season (Figure 2-5); t here were no statistically significant differences in seasonal cover within or between troops. Canopy cover varied by habitat, with the bamboo providing the most cover to both troops (Figure 2-6); there were no statistically significant differenc es in habitat cover within or between troops.
52 Group Size, Density and Variation in Travel Distance Troop A maintained an average troop dis per sion (mean SE) of 52.1 0.4 m2 with a density of 0.4 monkeys/m or 2.3 m/individual (not incl uding infants) within their daily range. Troop B maintained an aver age troop dispersion (mean SE) of 36.5 0.9 m2 with 0.2 monkeys/m or 4.1 m/indivi dual within their daily range. Troop A spent approximately 65% of their day trave ling (vs. stationary) and traveled, on average, a daily distance (mean SE) of 1,370 29 m (171 m/h). Troop B spent approximately 55% of their day trave ling (vs. stationary) and traveled a daily distance (mean SE) of 998 44 m (159 m/h). Time of day did not greatly impact either troops travel activity budget (Figure 2-7). Troop A traveled the most (70%) during the afternoon hours (2 to 5 p.m.) as compared to their morning (7 to 11 a.m.) travel (63%) and mid-day (11 a.m. to 2 p.m. ) travel (62%). Troop B traveled the most (68%) during the mid-day as compared to the morning (62%) and afternoon (55%). Both troops traveled the most during the transition season and the least during the wet season. Troop As daily travel rate (m/h) was significantly higher than Troop Bs daily travel rate ( Z = -2.09, P = 0.037, nA = 311, nB = 86). This difference, however, was seasonally based; Troop As daily travel rate was significantly higher than Troop Bs daily travel rate during the fruiting season (Z = 3.808, P < 0.0001, nA = 79, nB = 21), but not during the wet ( Z = -1.255, P = 0.210, nA = 87, nB = 28), dry (Z = -1.125, P = 0.261, nA = 88, nB = 27), or transition ( Z = -0.158, P = 0.874, nA = 57, nB = 10) seasons (Figure 2-8). Group Size and Variation in M aximiliana maripa Feeding During the months in which Maximiliana maripa fruited (February to July, with sporadic fruiting clusters in September and November), Troop A spent a total of
53 127.7 observed hours feeding on M. maripa (> 7% of total activity budget), with peak feeding rates in May and June (19% and 13% of monthly activity budget, respectively). On average ( SE), Troop A spent 4.4 1.3 min/h per month in a M. maripa tree. Troop B spent a total of 31. 2 observed hours feeding on M. maripa (> 8% of total activity budget), with a peak feeding rate in June (17% of monthly activity budget). On average ( SE), Troop B spent 5.7 1.2 min/h per month in a M. maripa tree. There was no significant difference between Troop A and Troop B in the monthly troop rate (min/h) of M. maripa feeding ( Z = -1.033, P = 0.345, nA = 8, nB = 6). After controlling for group size, however, Troop B individuals (adults and subadults) had a significantly higher monthly M. maripa feeding rate (mean SE = 1.26 0.27 min/h per individual) than did Troop A individuals (mean SE = 0.37 0.10 min/h per individual) (Z = 2.324, P =0.02, nA = 8, nB = 6). Group Size and Variation in Alarm Calls A Harpy Ea gle ( Harpia harpyja ) nest is located within the study site, and the nests residents are the major aerial predat ors to the primate population in RV (Boinski et al. in prep). The study period was a non-reproductive year for the Harpy Eagles, however, possibly because of extensive flooding that occurred during the wet season. The adult Harpy Eagles were still present in the area and observed to actively hunt, although at a reduced frequency compared to year s when nestlings were present (Boinski et al. in prep). A Harpy Eagle was observed near Troop A 5 times thoughout the year and detectable near Troop B on 2 occasions. No successful predation was observed. On average, Troop A produced AAs at a mo nthly rate (mean SE) of 0.62 0.06 alarms/h, while Troop B produced AAs at a ra te of 0.66 0.14 alarms/h. There was no significant difference between Troop A and Troop B in their monthly AA rates (Z =
54 -0.577, P = 0.590, nA = 12, nB = 12). After cont rolling for group size, however, Troop B individuals (adults, subadults and juveniles) produced significantly higher AA rates (AA/h per individual) t han Troop A individuals ( Z = 3.385, P =0.00, nA = 12, nB = 11) (Figure 2-9). The increased AA rate produced by Troop B individuals may have been a repercussion of height in canopy, as Troop B had a significantly higher upper height limit as compared to Troop A (mean SE = 12.23 0.29 m and 10.97 0.21 m, respectively) (Z = -3.510, P < 0.0001, nA = 228, nB = 78). There were no statistical intraor inter-group seasonal differences in AA rates; both troops produced AAs at the highest rate during the transition season (Fi gure 2-10). After controlling for group size, however, the average individual in Troop B produced AAs at a nearly significant ( P = 0.05) higher rate than did t he average (non-infant) indivi dual in Troop A during the fruiting, dry, and transition seasons but not during the wet season (Figure 2-11). Encounters between terrestrial predators and the C. apella study troops were infrequent, possibly due to the presence of human observers. Both Troop A and Troop B produced AAs at a significantly higher monthly rate than TPAs (Troop A: Z = -3.059, P = 0.002, n = 12; Troop B: Z = -3.059, P = 0.002, n = 12). There was one definite instance of a jaguar (Panthera onca ) within range of Troop A. The cat, however, already had a kill (undeterminable specie s) and did not appear interested in the capuchins. One jaguar attack on the troop wa s suspected, but cannot be confirmed as it occurred out of view in a stand of bamboo. There were no observed encounters between Troop B and a terre strial predator. Troop A produced significantly higher m onthly TPA rates (mean SE = 0.17 0.03 TPA/h) than Troop B (mean SE = 0.07 0.02 TPA/h) ( Z = -2.541, P = 0.01, nA =
55 12, nB = 12). After controlling for group size however, there was no significant difference between Troop As and Troop Bs individual rate of TPAs (TPA/h per individual) ( Z = -0.202, P = 0.843, nA = 12, nB = 12) (Figure 2-9). The increased TPA rate produced by Troop A may have been a repercussion of proximity to the ground, as Troop A had a significantly lower lower heigh t limit as compared to Troop B (mean SE = 4.95 0.16 m and 6.50 0.22 m, respectively) (Z = -5.727, P < 0.0001, nA = 228, nB = 78). There was an effect of season on group TPA rates: Both troops produced TPAs at the highest rate during the transit ion season, and Troop A produced TPAs at a (nearly) significant higher rate than Troop B during this time (Figure 2-12). After controlling for group size, however, there were no statistical intraor inter-group differences in individual TPA rates by season (Figure 2-13). Group Size and Variation in Female Agonism I recorded 875 agonis tic bouts among Troop A members (0.38 bouts/h), 676 of which (77%) were classified as decided bout s (i.e., a clear winner and loser were differentiated). Adult females were re sponsible for 137 bouts of agonism towards conspecifics (72% food-related), and fema les received 129 bouts of agonism (90% food-related) (Table 2-4). I recorded 41 agonistic bouts among Troop B members (0.08 bouts/h), 29 of which (71%) were classified as decided bouts. Adult females were responsible for 4 bouts of agonism toward s conspecifics (25% food-related), and females received 3 bouts of agonism (100% f ood-related) (Table 2-4). For both Troops A and B, there was no significant difference in the daily rate of intragroup agonism that females gave versus received (Troop A: Z = -0.755, P = 0.450, n = 300; Troop B: Z = -0.169, P = 0.866, n = 86). In comparing monthly rates of Troop A female agonism with rates of Troop B female agonism (after controlling for the number of females in
56 each group), Troop A females gave significant ly more agonism ( bouts/h per female) than Troop B females ( Z = -2.906, P = 0.004, nA = 12, nB = 12). Similarly, Troop A females received significantly more agonism (bouts/h per female) than Troop B females ( Z = -3.326, P = 0.001, nA = 12, nB = 12) (Figure 2-14). Within Troop A, agonism between females accounted for 6.8% of aggression, 19.5% of displacements, and 10.5% of s ubmission (Table 2-4). Food-related agonism comprised 82% of agonism, while 11% were of unknown context and 7% involved the alpha female threatening estrus fema les near the alpha male. Based on the outcomes of these agonistic bouts, the transitivity property, and anecdotal evidence (Chapter 3), it was possible to extrapolate a linear (although weak) dominance hierarchy for Troop A females (Table 2-5). The hierarchy could not be tested statistically because of small sample size (too few dyadic dominance interactions). Within Troop B, there was no observed agonism between females (Table 2-4). Therefore, it was not possible to derive a dominance hierarchy. Coalitions among Troop A me mbers that involved at least one female occurred at a rate of 0.02 coalitions/h. The primary targ et of these coalitions was a human observer (44.7%). Troop-mates were the second most frequent target of coa litions (38.3%), and coalitions involving the mobbing of another animal spec ies (e.g., tayra, tamandua) comprised 17% of coalitions. Coalitions wit hin Troop B that involved at least one female occurred at a rate of 0.07 coalitions/h. T he primary target of these coalitions was a human observer (87.9%). M obbing of another animal species comprised the remainder of the observed coalitions (12.1%); Troop B members were never observed to form a coalition against a troop-mate.
57 Group Size and Variation in Female Grooming I recorded 1574 dyadic groom ing bouts (with known dire ction and participants) between Troop A members (0.69 bouts/h). Of the 1013 bouts that involved at least one female (64% of troop grooming), a female acted as the groomer in 817 bouts (81%) for a total duration of 1937.7 min, and as t he recipient in 371 bouts (37%) for a total duration of 755.1 min. Female s in Troop A gave a significantly higher rate (mi n/h) of grooming than they received ( Z = -3.059, P = 0.002, n = 12). I recorded 363 dyadic grooming bouts (with known direction and participants) between Troop B members (0.73 bouts/h). Of the 202 bouts that involved at l east one female (56% of troop grooming), a female acted as the groomer in 141 bouts (70%) for a total duration of 303.6 min, and as the recipient in 93 bouts ( 46%) for a total duratio n of 143.9 min. Females in Troop B gave a significantly highe r rate (min/h) of grooming than they received ( Z = -2.353, P = 0.019, n = 12). Overall, Troop A had significantly higher da ily rates (min/h) of female grooming (given) than did Troop B ( Z = -3.448, P = 0.001, nA = 300, nB = 86), although there was no significant difference between the troops da ily rates of female grooming received ( Z = 1.423, P = 0.155, nA = 300, nB = 86). After controlling fo r the number of females in each group, however, there was no differ ence between Troop A and Troop B in the monthly rate (min/h per female ) of female grooming given ( Z = -0.462, P = 0.671, nA = 12, nB = 12) or received ( Z = -0.058, P = 0.977, nA = 12, nB = 12). In fact, although not statistically significant, Troop B fema les both gave and received higher monthly rates of grooming than did Troop A females (Figure 2-15). Further, of the grooming that occurred among adults ( and ), 51% of the dyads in Troop B were reciprocal, whereas only 30% of the adult grooming dyads in Troop A were reciprocal.
58 Group Size and Seasonal Variation in Social Behavior Troop A females had the highest average rate of grooming (given and receiv ed) during the fruiting season (February to April) the time of year that corresponded not only to fleshy fruit availability but also to the presence of young infants. While there was no significant seasonal variation in the rates at which females groomed troop-mates ( H = 4.443, df = 3, P = 0.217), there was significant seas onal variation in the rates at which females received grooming ( H = 20.089, df = 3, P < 0.0001). Specifically, females received more grooming during the fruiting ( Z = -4.161, P < 0.0001) and wet ( Z = -3.722, P < 0.0001) seasons than dur ing the dry season. Females in Troop A had the highest average rate of agonism (given and received) during the wet season (May to July), the time of year t hat corresponded to the peak fruiting of M. maripa In fact, during the wet season, 55% of the decided agonistic bouts involving at least one female occurred in a M. maripa tree. There was significant seasonal variation in the rates at which females gave ( H = 15.596, df = 3, P = 0.001) and received ( H = 23.843, df = 3, P < 0.0001) agonism. Females both gave and received significantly more agonism during the fruiting (Z = -2.480, P = 0.013; Z = -3.666, P < 0.0001, respectively) and wet ( Z = -3.468, P = 0.001; Z = -4.729, P < 0.0001, respectively) seasons than during the dry season, and females also gave ( Z = -2.551, P = 0.011) and received ( Z = -2.094, P = 0.036) significantly more agonism during the wet season than during the transition season. Regarding Troop B, there was no discer nable association between seasonality and variation in female social behavior. Only the rate at which females groomed conspecifics significantly varied by season ( H = 8.326, df = 3, P = 0.040): The average daily rate that a Troop B female groomed tr oop-mates was signif ic antly higher during
59 the transition season than during the wet ( Z = -2.529, P = 0.021) and dry ( Z = -2.405, P = 0.031) seasons, but this difference was due solely to an exaggerated grooming rate (12 times the average) performed by the troop s only remaining female during the month of November. Discussion The results of this chapter indicate that, based on the variables studied, the females in the smaller troop received more of the benefits of sociality, while the females in the larger troop incurred more of the cost s (Table 2-6). Below I discuss the outcome of each hypothesis, considering each variable of study and how it relates to the cost/benefit regime of social ity for brown capuchins in Raleighvallen, Suriname. Daily Travel Distance In creases w ith Group Size This hypothesis is supported and demonstrat es a cost of larger group size. Overall, the daily travel rate (m/h) was significant ly higher for Troop A than for Troop B. In response to increased WGC, a social group may increase t heir spread (thereby reducing the number of animals per unit of foraging space) or foraging effort in time and/or space (daily travel rate) (Snaith and Chapman 2007). Troop A employed both mechanisms, indicating a considerable effect of larger group size on the intensity of WGC. While Troop B maintai ned an average spread of 37 m2, Troop A increased their average spread to 52 m2. Despite this increase in ar ea, however, the intragroup density for Troop A was still two-fold t hat of Troop B. In terms of daily travel rate, Troop A traveled farther and faster and rested less than the smaller Troop B. In further support of the ecological-c onstraints model (Chapman and Chapman 2000), the difference between Tr oops A and B in their daily travel rate was most pronounced during the fruiting season when t he production of fleshy fruit was at its
60 peak. In RV, these fruit resources mostly occu r in small clumps (Boinski et al. 2002). In response to increased WGC that comes with larger group size, Troop A must search a larger area to ensure adequate nutriti onal intake for its members. In the dry season, however, Troop B actually demonstrated a higher daily travel rate than Troop A. During this time of y ear, fruit production is low, foliage cover is reduced (increasing the monkeys exposure to aerial predators) and the monkeys rely heavily on the protection and food resources (bamboo shoots and insects) provided by expansive, dense bamboo patches (Boinski et al. in prep). The importance of the bamboo patches is evident in the incr eased number and intensity of intergroup encounters that occur within this habitat during the dry season (Ehmke et al. in prep). The primary bamboo patch in the study site is approximately 200 m by 100 m and is centrally located within Troop As home range. It was not uncommon, especially during the dry season, for Troop A to enter this bamboo patch and remain within its perimeter for the entire day. The same bamboo patch is also accessible to Troop B along the southern border of their home ra nge, but they were often displaced from this resource by the larger Troop A and forced to travel in search of safety and/or food elsewhere. Thus, just as large-group females appear to be at a disadvantage during the fruiting season, small-group female C. apella in RV appear to receive a cost of reduced group size during the dry season. Both troops traveled the most during the transition season, a time of year when food resource availability is unpredictable and highly variable from year to year. Based on the similarity in the troops daily travel rates during this time, such unpredictability in food supply appears to lessen the effe ct of differential group size.
61 Individual Foraging Time in a Pr eferred, Monopolizable Food Tree ( M aximiliana maripa) Decreases with Increased Group Size This hypothesis is supported and demonstrat es a cost of larger group size. Overall, Troop B spent more time than Troop A feeding on M. maripa fruits, and Troop B individuals (adults and subadults) had a significantly higher monthly rate of M. maripa feeding than did Troop A adults and subadults. As with Hypothesis 1, this finding supports the ecological-constraints model su ch that individuals in a small group can feed in a single patch for longer periods as patches are more slowly depleted (Chapman and Chapman 2000). Predator Alarm Frequency (per individua l) is Inversel y Associated with Group Size I used alarm calls as a proxy measure of perceived predation risk. In an experimental study of wild baboons ( Papio cynocephalus ursinus ) investigating perceived predation risk (i.e., al arm call rates) as a function of group size, females in small groups appeared to perceive the highes t risk (Cowlishaw 1997b) In the current study, however, support for this hypothesis varies according to alarm call type and demonstrates a difference in perceived risk fr om aerial versus terrestrial predators according to group size. In terms of aer ial alarms, the hypothesis is supported and consistent with predictions t hat larger social groups benef it from overall levels of efficacy in predator detection even as individual effort decreases (Pulliam 1973; Bertram 1978). In terms of terrestrial predator alarms however, the hypothesis is not supported. For most diurnal primate species, it is commonly suggested that, controlling for body mass (Janson and Goldsmith 1995), group si ze should increase as predation risk increases (Dunbar 1988). So what explains such a disparity in group size within a single population? Effects of resource competition aside, divergent group size may be a
62 response to two different anti-predator strat egies. In RV, risk of predation from both aerial and terrestrial predators is strong. T he primary aerial threat to capuchin monkeys is the Harpy Eagle ( Harpia harpyja ), the largest raptor in the Americas (Brown and Amadon 1968). Harpy attacks on prey are ty pically accompanied by a series of vocalizations, and experienced prey are able to use this acoustic signal to initiate an escape response (Gil-da-Costa et al. 2003). Therefore, a small group may be better suited for protection from Harpy Eagles bec ause a cryptic, quiet group is harder to locate from the aerial view of a bird than is a large, conspicuous group, and the Harpy vocalizations somewhat preclude the need for many-eyes for early predator detection. On the other hand, a small group is likely more susceptible to quiet, ambush terrestrial predators, such as the jaguar ( Panthera onca ) and puma ( Puma concolor ) that rely heavily on scent for prey detection. The dense vegetation (in the form of lianas and bamboo) dominating RV is a double-edged sword: It provides a source of protection for the monkeys from aerial predators but also reduces visibility and provides camouflage to predatory cats, increasing the m onkeys risk of terrestrial predation. As potential prey for such terrestrial threats, a small capuchin troop is at a disadvantage due to their reduced group size (e.g., the dilution effect and many-eyes hypothesis). Thus while the larger Troop A may be more susceptible to attacks by Harpy Eagles, Troop B may be more susceptible to attacks by large cats. Disparate group size may therefore be a response to an env ironment with a strong predator population posing risk from both sky and ground. Evidence for this is provided by the troops habitat preference, range of height in the can opy (group strategy to reduce vulnerability) and alarm rates (individual response to t heir groups strategy). While bamboo
63 comprises a relatively equivalent portion of the troops home ranges, the dense bamboo patches were the habitat of choice for Tr oop A, while Troop B minimally occupied this habitat. Troop B stayed significantly higher in the canopy than did Troop A, and Troop B individuals rarely ventured to the forest floor, a patte rn that was found for small groups of other primate species as well, such as C. olivaceus (de Ruiter 1986) and Macaca fascicularis (van Schaik et al. 1983b). Overall, the data suggest that to reduce predation risk from Harpy Eagles, Troop A stayed lower in the canopy and in thick canopy cover; to reduce predation risk from terre strial cats, Troop B stayed higher in the canopy and avoided risky habitats. In response to these strategies, individuals in Troop B became more vigilant of aerial threats and produced more AAs than did the individuals in Troop A, while Troop A became more sensitive to terrestrial threats and produced more TPAs than Troop B. Perceived versus actual predation risk. For both troops, the frequency of alarm calls far surpassed the number of actual encounters with predators. In particular, although Harpy Eagle attacks on the study groups occurred relatively infrequently as compared to previous years, the monkeys perceived risk of predation was a strong indicator of past selective pressure (Isbell 1994). Differentiating between false and true alar ms (i.e., perceived vs. actual risk) requires identification and know ledge of the caller and involves opportunity cost (e.g., vigilance in lieu of foraging, resting or so cial behavior). In m any social birds and mammals, researchers report that false alarms represent a high proportion of all alarm calls (Beauchamp and Ruxton 2007); consist ent reaction to these false alarms can become additively costly in terms of time and energy expenditure. Using two different
64 mathematical models, Proctor et al (2001) and Beauchamp and Ruxton (2007) predicted how flocking birds should res pond to ambiguous alarm calls. Both demonstrated that, in general, birds should alwa ys respond to multiple alarms but never to single alarm calls, as single instances of alarm are likely false alarms. However, if flock size is small and/or attack time is short, the models predi ct that the most appropriate action is to respond to all alarms regardless of the number of calls. This may further explain the increas ed number of aerial alarm calls per individual in Troop B, many of which were secondary AAs in response to a conspecifics alarm. Rates of Female Agonism Increase with Group Size and Group Size Affec ts the Strength of Female Dominance Hierarchies These hypotheses are supported and demonstr ate a cost of larger group size. Among primates in female-philopatric so cieties with demonstrated within-group competition, females tend to form linear dom inance hierarchies, whereas in groups with weak within-group competition, females mainta in more egalitarian relationships (Sterck et al. 1997). The females in Troop A exhibited a weakly linear dominance hierarchy, while the relationship among the Troop B female s is best described as egalitarian; this further indicates that WG C was stronger within Troop A than it was within Troop B. Although the presence of a dominance hier archy likely acts to reduce aggression, females in Troop A still expressed higher ra tes of agonism (as actors and recipients) than the females in Troop B. This ma y be a repercussion of increased WGC, as approximately 80% of agonistic bouts involving females were food-related. Alternatively, increased agonism in Troop A may simply be a result of larger group size in that a greater density of individuals provides more occasion for (intentional and unintentional) interaction (Olsson and We stlund 2007). Based on group size (not
65 including infants), there were 504 possible dyadic associations within Troop A (45 possible adult dyads) and only 36 possible dyadic associations within Troop B (6 possible adult dyads). Rates of Female Groomi ng Increase w ith Group Size This hypothesis is not supported and demonst rates a cost of larger group size. Larger group size naturally provides a greater number of pot ential enemies and friends for any given individual (Olsson and Westlund 2007). The females in Troop A appeared to receive the negative social cons equences of increased group size (e.g., increased rates of agonism) but not the benefits (e.g., increased rates of grooming). Although the females in Tr oop A had double the density of neighbors and 14 times the number of possibly dyadic relationships t han Troop B females, the females in the smaller group were actually more invested in grooming. Given that Troop A females participated in grooming less than Troop B females, despite a threefold increase in troop-mates (and potential grooming par tners), it is of interest to understand what socioecological variables we re associated with female grooming and possibly acted to limit (or pr omote) grooming behavior. Is grooming a behavior that is better accommodated by t he home range use and travel patterns of Troop B? Posthoc consideration revealed t hat, for both Troops A and B, the two most common habitats in which grooming occu rred were the bamboo (40% and 17% of grooming bouts, respectively) and liana/low forest (31% and 64% of grooming bouts, respectively). These habitats were likel y preferred by the monkeys for grooming because they provided the most canopy cove r from aerial predators. That Troop A groomed most frequently in the bamboo and Tr oop B clearly preferred the liana/low forest for grooming is likely a reflection of their habitat use patterns in response to their
66 respective sources of predation risk (as discussed above). Still, however, the home range and habitat use patterns of the two tro ops does not explain why Troop A groomed so much less than expected; Troop A actually spent a greater percentage of their time (66%) in the habitats amenable to groomin g as compared to Troop B (64%). A more likely explanation is provided by the groups activity patterns. Pollard and Blumstein (2008) demonstrated t hat an evolutionary relationship exists between time allocation to rest and social group size. A ccording to their meta-analysis, time spent resting was the most important time component of group size such that larger groups allocated less time to rest than did smalle r groups. In my study, the females in both troops participated in grooming more than ex pected when the group was stationary as opposed to traveling (Troop A: X2 = 53.2, df =1, P < 0.001; Troop B: X2 = 86.2, df = 1, P < 0.001). Troop A traveled more than Troop B on a daily basis, and the travel budget of Troop A was approximately 10% higher than that of Troop B. Therefore, just as smaller groups are able to allocate more time to resting, Troop B remained more stationary and were provided with more opportuni ty to groom as compared to the larger Troop A. Decreased grooming time is an opport unity cost of residing in a large group that must travel more in or der to feed its many mouths. Female Brown Capuchins in a Larger Troop Demonstrate More Seasonal Variation in Patterns of Grooming and Ag onism than Do the Females in a Smaller Troop This hypothesis is supported and demonstrates a cost of larger group size. While individuals in a small group are more likel y to experience social stability in an ecologically unstable environment, the same small-group individuals, however, are likely more vulnerable to large-scale disruption, es pecially over the long term. For example, the loss of one of t he Troop A females during the study provided no measured reaction
67 by the remaining group mem bers; the large group seemed to act as a buffer to this individual loss. Troop B, however, simultan eously lost two of its three adult females, and troop reaction was more strongly observed. Anecdotal evidence suggested that the alpha male demonstrated an intense behavioral response in the days following their disappearance, and although I cannot attribute a causal link with any degree of certainty, it is of interest that the remaining female unc haracteristically intensified grooming behavior during the transition seas on that followed. Therefore, barring large scale disruption, a small, cohesive group provides stability and likely circumvents the effects of daily and seas onally predictable stress. During large-scale social or ecological upheaval, however, a large so cial group may buffer the effects of unpredictable stress. Conclusion It is often stated that group size in wild populations is self-regulated as indiv iduals choose to stay in or migrate from a so cial group depending on th e benefits obtained. For the females in a female-philopatric societ y, however, the situation is typically more of a given than a choice. For these fema les, philopatry can be viewed as a subtle version of captivity. With the i mminent costs associat ed with emigration and immigration (increased predation risk, lack of co mpetitive ability in acquiring resources, lack of social ties and allies, and risk of inf anticide) and the benefits provided by a social network founded in kinship and long-time familia rity, a female is almost certain to choose to remain in her natal group regardless of its size (but se e Izar 2004). She is born into a situation where she must endure the challenges of her group. Depending on the local ecology of the environment, as well as intraand inter-group dynamics, there are inherent trade-offs to liv ing in a small versus a large group (and vice versa).
68 In this chapter, I tested six hypothe ses addressing the costs and benefits of disparate group size. If a group of intermedi ate size provides, in theory, the best balance of between-group and withi n-group competition (Price and Stoinksi 2007), then how does one go about explaining the successful history (eight y ears and counting) of these two troops of brown capuchins in RV, Suriname that represent completely opposite ends of t he spectrum of C. apella group size? The answer is provided by the dynamic nature of their local ecology. With intense ecological variation, both within and between years, and a st rong regime of aerial and terrest rial predators, deviation from the norm is most adaptive. When the cost/benef it scale can easily shift, what is best for a small group at one point in time may soon be better for a large group a cycle of unpredictability that has maintained a dr astic differential in group size. The study period proved to be a beneficial time for the individuals in the small group, Troop B. Based on the variables studied, Troop B members received more of the benefits of sociality while Troop A mem bers received most of the costs. Troop A must travel more in order to acquire su fficient nutrients for its many group members, and because their large size makes them more conspicuous to aerial predators, Troop A tended to spend most of their time in the densely covered bamboo patches. Only during the fruiting season, when fleshy fruits reach peak production, was Troop A lured from the protective bamboo. Troop B was not as constrained by the threat of aerial predation, however, and they ut ilized bamboo only as a fallback resource. And while the females in Troop A were exposed to higher rates of agonism (i.e ., WGC) from their many troop-mates, they were less invested in grooming than were the females in Troop B. This low rate of grooming in Troop A is of particular interest given the number of
69 potential grooming partners, t he presence of infants (which tends to increase female attractiveness to grooming conspecifics (OBrien 1993; Muroyama 1994; Perry 1996; Di Bitetti 1997; Manson et al. 1999; Henzi and Barrett 2002), their access to safe grooming habitat (primarily the liana/low forest and bamboo), and the presence of a female dominance hierarchy. I examine the fine-scaled grooming relationships of the females in Troop A in greater detail in Chapters 4 and 5.
70 Table 2-1. Reported (or calculated mean) group size of C. apella troops at various sites. The two study groups, Troops A and B, represent opposite ends of the continuum of C. apella group size. Location X group size Source El Tuparro, Colombia 10 Defler 1982 Manu, Peru 12 Janson 1990b Manu, Peru 13 Janson 1985 Manu, Peru 13 Terborgh 1983 Atlantic Forest, Brazil 13 Izar 2004 Nouragues, French Guiana 13 Zhang and Wang 2000 Manu, Peru 15 Janson 1996 Iguazu, Argentina 15 Di Bitetti and Janson 2001 Nouragues, French Guiana 15 Zhang and Wang 1995 La Macarena, Colombia 16 Izawa 1980 Iguazu, Argentina 18 Di Bitetti 1997 Iguazu, Argentina 25 Hirsch 2002 Caratinga, Brazil 26 Lynch et al. 2002 Optimal (intermediate) C. apella group size 16 RV, Suriname: Troop A RV, Suriname: Troop B 27 9 Current study Table 2-2. Composition of the study groups, Troop A and Troop B, from January to December 2006. Parentheses indicate final troop composition at the conclusion of the study. Troop A lost one female in October (death presumed to be related to giving birth) and two subadult males in August (emigrated). Troop B lost two females simultaneous ly in August (unknown cause). Group size # adult # adult # subadult # subadult # juv # juv # infants (<1yr) Troop A 27 (24) 3 7 (6) 2 (0) 1 7 3 4 Troop B 9 (7) 1 3 (1) 1 0 3 1 0
71 Table 2-3. Operational definitions for female C. apella agonistic and grooming behavior, recorded ad libitum Agonism Aggressive Displacement Submissive All aggressive, displacement, and submissive behaviors Actively antagonistic behaviors such as chase, hit, bite, and threat display. One individual passively suppl anted by another, usually (but not exclusively) at a food source. If an individual aggressively displaced another individual, then the bout was scored as aggression and not displacement. Any instance in which an indivi dual retreated and/or reacted with acquiescent posture or vocalizations without being aggressed towards or displaced. Grooming bout One individual picking through the hair and/or skin of another individual. The bout conclude s when the grooming ends or is interrupted by another behavior for more than five seconds. A new grooming bout begins with a new dy ad, or whenever the grooming resumes (after 5 seconds) or changes direction within the same dyad.
72 Table 2-4. Summary of the observed intragroup agonistic bouts involving females (with determinable winner and loser) for Troop A and Troop B. The females in Troop B rarely experienced agonism, while agonistic bouts occurred more frequently for females in Troop A. Troop A female s were most frequently aggressive towards immatures and received the most aggression from adult males. Agonism between the females in Troop A occurred mostly in the form of passivity; agonsim between females in Troop B was never observed. Actor Recipient Troop A Troop B Aggression Displacement Submissive Aggression Displacement Submissive Female Female Female Male Juvenile* Male Juvenile* Female Female Female 6 59 8 41 4 1 23 16 30 12 0 20 4 11 3 0 0 0 1 2 0 0 0 0 0 0 4 0 0 0 Notes: Aggression is action taken by the actor whereas Displacement and Submissive behaviors are actions taken by the recipient without obvious provocation. Female and Male denotes adult indivi duals of the respective sex; Juvenile includes all immature individuals of both sexes. Table 2-5. Troop A female dominance matrix based on the outcomes of dyadic agonism. Gina Banana Jane LH Carol Kate Tinkerbell Gina 1 1 1 ^1 1 1 Banana *1 1 1 1 ^1 Jane 1 1 ^1 1 LH 1 ^1 *1 Carol 1 *1 Kate *1 Tinkerbell Note: 1 = row individual is dominant to column individual, based on the outcome of agonistic episodes. ^1 = dominance decided based on transitivity. *1 = dominance decided bas ed on anecdotal evidence.
73 Table 2-6. Differential benefits and costs for individual female brown capuchins residing in a large group (Troop A) versus a sm all group (Troop B) in Raleighvallen, Suriname. Based on the variables studied, Troop B members received more of the benefits of sociality while Troop A members received more of the costs. In terms of the benefits of sociality (represented by check marks), individual females in Troop B experienced greater social cohesion (as indicated by higher rates of grooming, grooming reciprocity, and agonistic coalitions), while members of Troop A were able to be less vigilant of aerial predators. The large group size of Tro op A likely provides long-te rm social stability, albeit less support during daily challenges; t he smaller Troop B provides a more supportive system overall, but is more susceptible to large-scale disturbances (such as the loss of troop members). In terms of the costs of sociality (represented by X marks), Troop A females experienced increased withingroup food competition (as indicated by increased daily travel distance, presence of dominance hierarchy, and dec reased foraging time per individual in clumped food patches) and increased intragroup agonism. Troop A (large group) Troop B (small group) Benefits of Sociality Reduced predation risk < AAs/individual < TPAs/individual Social cohesion > grooming rates > grooming reciprocity > coalitions Social stability Long-term Short-term Costs of Sociality Increased WGC > daily travel distance > dominance hierarchy < time/individual in clumped food patches Increased intragroup agonism
74 10% 44% 43% 3% Bamboo High Forest Liana/low forest Swamp Figure 2-1. Habitat composition of t he home ranges of Troop A and Troop B. Troop As home range was composed primarily of high forest and secondarily of liana/low forest, while Troop Bs home range was equally dominated by high forest and liana/low forest. 13% 54% 28% 5%Troop A Troop B
75 Figure 2-2. Comparison of overall time (%) that Troop A and Troop B spent in each habitat type. Overall, Troop A spent t he most time in the bamboo while Troop B spent the most time in liana/low forest. Troop A spent significantly more time in the bamboo than did Troop B ( X2 = 177.88. P < 0.001), while Troop B spent significantly more time in th e liana/low forest than did Troop A ( X2 = 63.86, P < 0.001). 0 10 20 30 40 50 60 BambooHigh forestLiana/low forest Swamp Troop A Troop B** P < 0.001 ** P < 0.001 Time ( % )
76 Figure 2-3. Seasonal comparison of Troop As habitat use (% time). In each season except the fruiting season, the group spent the most time in the bamboo, especially during the dry season. Duri ng the fruiting season, the group spent the most time in the high fo rest. Their use of the liana/low forest, especially in conjunction with the high forest was relatively consistent. Figure 2-4. Seasonal comparison of Troop Bs habitat use (% time). In each season except the transition seas on, the group spent the most time in the liana/low forest. Their use of the high forest remained relatively consistent throughout the seasons while their use of the bam boo was more variable, peaking in the transition season. 0 10 20 30 40 50 60 FruitingWetDryTransition Bamboo High forest Liana/low forest Swamp 0 10 20 30 40 50 60 70 FruitingWetDryTransition Bamboo High forest Liana/low forest SwampTime ( % ) (Feb-Apr) (May-Jul) (Aug-Oct) (Nov-Jan) Time ( % ) (Feb-Apr) (May-Jul) (Aug-Oct) (Nov-Jan)
77 Figure 2-5. Seasonal comparison of average canopy cover (0-3 incremental scale that increases with vegetation density) for Troop A and Troop B. The two troops demonstrated a similar cover score to one another in each season except the dry season. The cover differentia l between Troop A and Troop B during the dry season, however, was only 0.2 on a three-point scale. There were no statistically significant interor intr a-group differences in seasonal cover. Figure 2-6. Average canopy cover by habi tat type for Troop A and Troop B. Both troops were provided the most cover by the bamboo, a condition that reduced exposure to aerial predators but also reduced visibility between conspecifics and increased susceptibility to terrestr ial ambush predators. There were no statistically significant interor intr a-group differences in habitat cover. 2.15 2.2 2.25 2.3 2.35 2.4 2.45 2.5 2.55 FruitingWetDryTransition Troop A Troop B 0.0 0.5 1.0 1.5 2.0 2.5 3.0 BambooHigh forestLiana/low forestSwamp Troop A Troop B(Feb-Apr) (May-Jul) (Aug-Oct) (Nov-Jan) X cano py cover SE X cano py cover SE
78 Figure 2-7. Percentage of time Troops A and B remained stationary (vs. travel) during the morning (7:00 to 11:00 a.m.), midday (11:00 to 2:00), and afternoon (2:00 to 5:00 p.m.) hours. During each time period, Troop B was more stationary than Troop A. Troop A travel ed the most (70%) during the afternoon hours as compared to the morning (63%) and mi d-day (62%), while Troop B traveled the most (68%) during the mid-day as compared to the morning (62%) and afternoon (55%). Figure 2-8. Seasonal comparison of daily tr avel rates (m/h SE) for Troops A and B. Overall, the daily travel rate of troop A was significantly higher than that of Troop B ( P = 0.037), but the difference was primarily based on the difference in their travel rates dur ing the fruiting season ( P < 0.0001). 0 10 20 30 40 50 60 MorningMid-dayAfternoon Troop A Troop B 0 50 100 150 200 250 FruitingWet DryTransition Troop A Troop B 7-11 a.m. 11 a.m -2 p.m. 2-5 p.m. Time ( % ) s p ent stationar y X dail y travel rate (Feb-Apr) (May-Jul) (Aug-Oct) (Nov-Jan)
79 Figure 2-9. Monthly group and in dividual rates (alarms/h SE) of aerial alarms (AA) and terrestrial predator alarms (TPA) for Troops A and B. Overall, there was no statistical difference in the monthl y rate at which Troop A and Troop B produced AAs, but Troop A produced signifi cantly more TPAs than did Troop B. After controlling for group size however, Troop B individuals produced AAs at a significantly higher rate than did Troop A individuals while there was no statistical difference in the rate at which Troop A and Troop B individuals produced TPAs. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 AATPAAATPA Group rate Individual rate Troop A Troop B** (P = 0.01) ** (P = 0.00) X monthly alarm rates (alarms/h)
80 Figure 2-10. Seasonal effect on group rates of aerial alarms (AAs/h SE) for Troops A ( n = 23) and B ( n = 9). Both troops produced t he highest rate of AAs during the transition season, although there were no intraor inter-group statistical differences in seasonal AA rate. Figure 2-11. Seasonal effect on the rates of aerial alarms (AAs/h SE) for individuals in Troop A and B. Controlling for troop size, an individual (juvenile, subadult, or adult) in Troop B produced AAs at a nearly significant higher monthly rate (AAs/h) than did an individual in Troop A during the fruiting, dry, and transition seasons. 0 0.5 1 1.5 2 FruitingWetDryTransition Troop A Troop B 0 0.05 0.1 0.15 0.2 0.25 FruitingWetDryTransition Troop A Troop B(Feb-Apr) (May-Jul) (Aug-Oct) (Nov-Jan) X monthly aerial alarm rates by season (P = 0.05) (P = 0.05) (P = 0.05) X monthly aerial alarm rates by season (Feb-Apr) (May-Jul) (Aug-Oct) (Nov-Jan)
81 Figure 2-12. Seasonal effect on the group ra tes of terrestrial pr edator alarms (TPAs) (TPAs/h SE) for Troops A ( n = 23) and B ( n = 9). Both troops produced the highest rate of TPAs during the transiti on season. During all seasons, Troop A emitted a higher group rate of TPAs than did Troop B; the difference, however, was (nearly) statistically si gnificant only during the transition season. Figure 2-13. Seasonal effect on the rates of terrestrial predator alarms (TPAs/h SE) for individuals in Troop A and B. Controlling for troop size, an individual (juvenile, subadult, or adult) in Tr oop B produced TPAs at a higher (nonstatistically significant) rate than did a non-infant individual in Troop A during the fruiting, wet, and dry seasons but not during the transition season. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 FruitingWetDryTransition Troop A Troop B 0 0.005 0.01 0.015 0.02 0.025 FruitingWetDryTransition Troop A Troop B(Feb-Apr) (May-Jul) (Aug-Oct) (Nov-Jan) X monthly terrestrial predator alarm rates by season (Feb-Apr) (May-Jul) (Aug-Oct) (Nov-Jan) X monthly terrestrial predator alarm rates by season
82 Figure 2-14. Controlling for group size, the monthly rate (bouts/h SE) at which the average female in Troop A and Troop B experienced agonism (as actor vs. recipient) with troop-mates. The majority of agonism was food-related. The average female in Troop A gave ( P = 0.004) and received ( P = 0.001) significantly more agonism than did the average female in Troop B. Figure 2-15. Monthly rate (mi n/h SE) that the average female in Troop A and Troop B acted as the groomer and as the reci pient during grooming. Although not statistically significant, females in Troop B tended to groom and be groomed more than the females in Troop A despi te the threefold increase in the number of potential gr ooming partners for females in Troop A. 0 0.002 0.004 0.006 0.008 0.01 0.012 actor recipient Troop A Troop B 0 0.05 0.1 0.15 0.2 0.25 0.3 groomer recipient Troop A Troop B** (P = 0.004) ** (P = 0.001) Monthly rate of agonism Monthly rate of grooming
83 CHAPTER 3 DISPARATE TROOP SIZE AND STRESS: THE EFF ECTS OF A HIGH RISK, LOW ENERGY ENVIRONMENT ON FEMALE BROWN CAPUCHINS Introduction Group size is a critical, though relatively dynamic component of social living a demographic feature dependent on local ecology and social circumstances that is associated with potentially severe fitness consequences. The effects of group size not only have direct and immediate fitness cons equences on individuals (e.g., predation risk and access to mates and food resources), bu t may also indirectly affect individual fitness in terms of physical development and reproductive health. With the benefits of sociality come the costs of group living, and suboptimal group size can exacerbate socioecological stress incurred by individuals. Females in femalephilopatric societies are especially impacted by the socioecological and endocrinological consequences of group size, as they tend to remain in t heir natal troop regardle ss of its size. Cortisol, a steroid hormone, is commonly measured as an indicator of stress in wild animal populations (Wingfield et al. 1997). Over the short term, acute stress response is an adaptive physiological condit ion whereupon energy is diverted from long-term processes such as digestion, growth, immune function and reproduction and mobilized to meet short-term needs (flight or fight), thus increasing an individuals likelihood of immediate survival upon encount ering a threat. Over extended periods, however, the metabolic cost s of sustained stress rise steeply (Sapolsky et al. 2000), adversely affecting female health and repr oductive ability, fetal development, and the development of offspring social skills and st ress-coping abilities (Sachser et al. 1998; Bardi et al. 2005; Sapolsky 2005; Shively et al. 2005; Wadhwa 2005). Fecal hormones are pooled as they metabolize for a period of time (usually less than 48 h in nonhuman
84 primates) before elimination (Whitten et al. 1998) and do not capture minute hormonal fluctuations. Therefore, fecal corticosteroid metabolites provide a useful indicator of pronounced or chronic stress (Boinski et al. 1999b). This chapter investigates the effect of group size on the intensity of socioecological stress incurred by resident females. A population of brown capuchins ( Cebus apella ) in Raleighvallen (RV), Suriname provides a unique opportunity to investigate this relationship: Two study groups are of widely disparate sizes ( n1 = 27, n2 = 9) but reside with ov erlapping home ranges withi n a high-risk, low energy environment. My study is the first to exami ne social and ecological stressors of wild female brown capuchins; published reports of Cebus behavioral endocrinology are scant and limited to C. apella males, captive C. apella females and wild C. capucinus females. Further, within Cebus female behavioral endocrinology has been studied only in terms of sexual behavior and reproductive physiology (Linn et al. 1995; Carosi et al. 1999; Carosi and Visalberghi 200 2; Carnegie et al. 2005). Hypotheses Using a combined approach of behavioral observation and hormonal analysis, my study focuses on three hypotheses investigat ing the effect of disparate group size on the socioec ological stress incurred by the groups resident females. Group Size Affects the Level of Stress Females Incur Due to Predation Risk Predation risk is accepted to be one of the ultimate causes of sociality (van Schaik 1983) and thus is a determinant of group size. Encounters with predators or their cues activate the hypothalamic-pituitary-adrenal (HPA) axis to release epinephrine and glucocortic oids, hormones that prepare and protect the body dur ing stress response (Boinski et al. 1999a; Engh et al. 2006a) Engh et al. (2006a) found that the
85 glucocorticoid levels of female chacma baboons ( Papio hamadryas ursinus ) increased significantly following a predation event, and Boinski et al. (1999a) demonstrate a positive correlation between the alarm rates of captive brown capuchins and fecal cortisol levels. High cortisol levels have also been found to occur in avian populations facing high predation risk (Scheuerlein et al. 2001). If predation risk is strong enough within a population to promote larger group size (as ex hibited within RV by Troop A), then animals facing high predator pressure (those individuals in the same population residing in smaller groups, such as Troop B) should have high cortisol levels (Pride 2005b). I predicted that the fe males in the smaller group would be more susceptible to increased cortisol levels in response to predation risk. Group Size Affects the Levels of Str ess Females Incur During Food Sh ortages Prior studies on various social taxa, e.g. baboons (Sapolsky 1986; Gesquiere et al. 2008), African elephants (Foley et al. 2001), ma le brown capuchins (Lynch et al. 2002), and ring-tailed lemurs (Pride 2005a), demonstr ate increased cortisol levels during periods of harsh environmental conditions and low food availability. Is there a contributing effect of group size on an i ndividuals stress response to cyclical environmental challenges? Wrangham (1980) suggested that female phil opatry is a response to intergroup competition; by this measur e, large group size should be preferred over small group size, as larger groups easily supplant small groups during competitiv e encounters. For most taxa, however, within-gr oup food competition has a great er effect on the dynamics of female relationships than does betw een-group competition (Janson 1988a; Cheney 1992; Barton and Whiten 1993; Cowlishaw 1995; S ilk et al. 1999; Koenig 2000; Boinski et al. 2002), and thus individuals in smaller gr oups may fare better in terms of access to
86 resources than do individuals in larger groups. Hormonal measures further indicate the significance of within-group competition fo r group-living species. Pride (2005a) demonstrated that, during periods of food scarc ity, cortisol levels increased for ringtailed lemurs only in large groups. Similarl y, Foley et al. (2001) determined that, for African elephants, cortisol levels increased with group size during the dry season. Therefore, I predicted t hat (a) the cortisol levels of fe males would be highest during the dry season (August to October), and (b) during the dry season, mean cortisol concentrations would increase more for fe males in the large group as compared to females in the small group. Rank-Related Stress is a Factor of Group Size Although rank-related stress is a dominant theme in t he study of behavioral endocrinology of group-living ani mals, there remains no consensus on the rel ationship between social status and cortisol level. Despite the original assumption that subordinance induces stress, only a few studies demonstrate that subordinate individuals have higher glucocorticoid le vels (Manogue et al. 1975; Virgin and Sapolsky 1997). Other studies demonstrate the opposite trend, with dominant individuals having higher stress levels than subordinates (Saltz man et al. 1998; Cavigelli 1999; Barrett et al. 2002; Cavigelli et al. 2003). A third trend in rank-related stress has actually received the most support, that of no consistent di rectional relationship between dominance ranking and stress levels (van Schaik et al. 1991; Bercovitch and Clarke 1995; Smith and French 1997; Stavisky et al. 2001; Lynch et al. 2002; Weingrill et al. 2004; Gould et al. 2005; Peel et al. 2005; Pride 2005a). Predictable stressors induce less of a stress response than do unpredictable and uncontrollable situations (Wi ngfield et al. 1998), and since a stable social system
87 produces predictable relationships with dominance interactions that do not require overt aggression, hierarchical position does not necessarily induce endocrine response (Sascher et al. 1998). Weingrill et al. ( 2004) and Pride (2005b) dem onstrated that in female chacma baboons and ring-tailed lemurs respectively, social rank was not a predictor of cortisol level; for both specie s the social hierarchies are stable, usually uncontested for long periods of time, and were characterized by little overt aggression. Similarly, in brown capuchins, females may exhibit a weak-ranked hierarchy (Izar 2004) and even mild overt aggression between C. apella females at RV is rare. Even in the absence of a consistent link between stress and social rank, however, does disparate group size differently affect i ndividuals of disparate ra nk? If group size affects the intensity of within-group resource competition, then it follows that larger groups may demonstrate a more linear dominance hierarchy with subordinate individuals more susceptible to stress th rough increased aggression and access to only lower quality food resources. Th erefore, I predicted that the cortisol levels of dominant and subordinate females would di ffer more in the large group than in the small group. Methodology Site Description Raleighvallen (RV), a 7812 km2 reserve consisting of primary tropical rain forest, is located within the 1.6 million ha Central Suri name Nature Reserve in Suriname. The main study site measured approximately 2 km2 and was covered by an extensive trail system. RVs flora and fauna are effectively undisturbed in the historical period, and an intact array of potential predators and competitors are present (Reichart 1993). The soils of the Guyanan Shield (including French Guiana, Suriname and Guyana) are nutrient-poor and highly weathered as they are derived from Precambrian bedrock.
88 Fleshy fruits are low in abundance and available fruit resources typically occur in small patches (< 5 m diameter) (Boin ski et al. 2002). Additionally, much of RV is composed of expansive bamboo patches and dense liana forest t hat restricts visibilit y (Boinski et al. 2003). Study Animal The brown capuchin monkey ( Cebus apella ) is one of the four traditionally recognized species in the genus Cebus (Hill 1960). Capuchins li v e in multi-male, multifemale polygamous groups that are typically female philopatric. Adult males and females maintain separate dominance hierar chies; in general, males are dominant to females. C. apella social groups are comprised of 12 to 27 individuals, with an average troop size of 17 members. Troop membersh ip and relationships remain relatively stable, except for subadult males who em igrate from their natal group and may subsequently transfer several times. Brown capuchins reside in a variety of habitats encompassing a large geographic range from Columbia to Argentina (Fragaszy et al. 2004). C. apella are omnivores; the bulk of their diet consists of fruit supplemented with seeds, vegetation, arthropods, and vertebrates (Terborgh 1983). Study troops A and B were two of f our brown capuchin groups commonly observed within the study site, and the primar y study subjects were the 10 adult females resident to these two groups. For most of the study, Troop A consisted of 27 individuals, seven of which were adult fema les; Troop B consisted of nine individuals, three of which were adult females. Duri ng the study, one pregnant female disappeared from Troop A (her disappearance coincided with her predicted du e date, and thus it is possible her (presumed) death was rela ted to giving birth), and two females disappeared from Troop B (cause unknow n). Relatedness among the adult troop
89 members is unknown; however, genetic analysi s of the brown capuchin population in RV is currently underway. The study tr oops had been habituated to human observers since 1998, and all individuals were recognized based on body size, color patterns, and other identifying characteristics such as scars, moles, and ear shape. Throughout the study period, inter-observer reliability tests were conducted monthly to ensure consistent and accurate identif ication of all troop members. Data Collection From January to December 2006, I and/or fi eld assistants followed Troop A for up to 10 hours each day (beginning at sunris e). Troop A was followed, on average, 28 days per month, composing a total of 2284 h of observation (not including out of view time). Troop B was followed, on average, 8 days per month, composing a total of 495 h of observation (not includ ing out of view time). Behavioral data, including all observed occurrences of social agonis m and anti-predator behavior, were collected ad libitum (Altmann 1974). To reduce observation bias, I moved throughout the gr oups to locate non-visible individuals and record their activities. Due to opportunistic sampling, as well as the reduced visibility within the Raleighvallen forest, the reco rded frequencies of social behavior are likely underestimates of the true frequencies. Bec ause of the large number of hours over which behavioral interactions were recor ded, however, I am confident that the data provide strong indication of actual trends. Hormonal Sampling and Analysis Fecal samples were collected opportunistica lly, with the goal of collecting at least one sample per femal e ( n1 = 7, n2 = 3) per 15-day interval (Lynch et al. 2002; Engh et al. 2006a). Fecal samples were collected only if the female was observed defecating and
90 her identity ascertained with certainty. U pon collection, each sample was immediately placed in an empty plastic vial and stored inside a thermos with an instant chemical cold pack. Samples were processed within th ree hours of collectio n and registered, on average, at 44F. The samples were proc essed using Solid Phase Extraction (SPE) according to Ziegler and Wittwer (2005): 0.1 g of fecal materi al was mixed with 2.5 ml of distilled water and 2.5 ml of et hanol. The mixture was then hand shaken for 5 min and centrifuged for 10 min. Two ml of the s upernatant was removed and passed through an Alltech Prevail C18 Maxi-Clean SPE Cartridge (Alltech, Deerfield, IL) and stored. Samples were shipped to The Wisconsin Regional Primate Res earch Center for analysis. Using Enzyme ImmunoAssay, eac h sample was analyzed for cortisol (F) concentration (ng/g); the assays were va lidated for accuracy and parallelism using internal controls. Statistical Analysis Using the individual female as a sampli ng unit allowed me to assess variation within eac h group as well as between groups. Th is is justified by ev olutionary theory, as selection is generally strongest at the level of the individual, and biologically sensible, because individuals may respond differently to the same environmental stimuli (Pride 2005a). A baseline fecal cortisol level for eac h female was established by calculating the mean after removing the highest and lowest 5% of values (Peel et al. 2005). These baseline cortisol levels were used when co mparing the two troops female cortisol concentrations and for establishing the relationship between cortisol and state (long term) variables, such as female dominanc e rank and seasonal food availability. When establishing the relationship betw een female cortisol levels and event
91 (short term) variables, such as rates of alar m calls and social agonism, however, all cortisol values ( nTroop A = 265, nTroop B = 42) were included in analysis. Non-parametric analyses were used because data failed to meet the assumption of normality. Mann-Whitney U and Kruskal-Wallis tests were performed using SPSS 11.0. Correlations between parameter s were evaluated using Spearmans rs (Excels Analyse-It statistical package). The significance level for all tests (two-tailed) was set at 0.05. Mean values are presented with standar d error values, unless otherwise noted. Results Fecal Cortisol Profiles A total of 265 fecal samples were colle cted from Troop A females, and 42 samples were collec ted from Troop B females. Baseline fecal cortisol concentrations ranged from 28.75 to 75.41 ng/g for t he seven Troop A females ( n = 237) and from 25.34 to 48.58 ng/g for the three Troop B females ( n = 34) (Table 3-1). Troop A females had, on average, higher baseline cortisol concentra tions (mean SE = 70.45 4.58 ng/g) than Troop B females (mean SE = 41.13 6.40 ng/g), but the difference was not statistically significant ( Z = 0.442, P = 0.658). Over the course of the year, the two groups demonstrated similar pat terns of monthly variation in their females baseline cortisol profiles (Figure 3-1). Diurnal Variation in Female Cortisol Levels While time of day affects plasma and ur inary cortisol concentrations (Coe and Levine 1995; Sousa and Ziegler 1998), this effect is less evident in measures of fecal cortisol due to the 1 to 2 day accumulati on time in hormone content (Campbell et al. 2001). In my study, there was no discernable relationship between troop-level female baseline cortisol values and time of day. No significant difference existed between
92 morning versus afternoon female cortisol concentrations for either Troop A ( Z = -0.401, n = 237, P = 0.688) or Troop B ( Z = -0.248, n = 34, P = 0.804), and while the females in Troop A demonstrated both higher morning and afternoon averages than the Troop B females, the difference was not statistically different ( Zmorning = -0.435, n = 98, P = 0.664; Zafternoon= 0.237, n = 173, P = 0.813) (Figure 3-2). Furt her, when opportunity allowed, I collected multiple samples per female per day. Evaluation of these samples revealed no consistent increase or decrease in cortisol with time. Similarly, there was no discernable pattern between diurnal variation in individual female cortisol concentration and seasonalit y. Only one Troop A female, Gina, had a significantly higher mean morning F value ( Z = -2.854, n = 29, P = 0.003), and the difference was significant only in the dry season ( Z = -2.191, n = 14, P = 0.022). On the other hand, one Troop A female, Tinkerbell, had a nearly significant higher mean F value in the afternoon than in the morning ( Z = -1.907, n = 32, P = 0.058), and this difference approached significance only in the fruiting season ( Z = -2.121, n = 7, P = 0.057). Troop Size, Female Cortisol Levels and Predation Risk As discuss ed in Chapter 2, after correcti ng for differences in troop size, Troop B individuals produced significantly higher mont hly rates of aerial alarms (AA) (mean SE = 0.09 0.02 AA/h per individual) than Troop A individuals (mean SE = 0.03 0.003 AA/h per individual) ( Z = -3.385, P = 0.00, nA = 12, nB = 11). No correlation existed, however, between the monthly AA ra tes of Troop B and the females monthly F levels ( rs = -0.30, P = 0.431), while a strong, positive correlation existed between the monthly AA rates of Troop A and t he females monthly F levels ( rs = 0.74, P = 0.009). Further, an even stronger correlation ex isted between Troop As monthly female F
93 levels and the monthly frequency with which a Harpy Eagle was observed to attack and/or be in proximity (within a 50 m radius) to the group ( rs = 0.80, P = 0.002) (Figure 3-3), while no correlation resulted between the same variables for Troop B ( rs = -0.25, P = 0.517). As was shown in Chapter 2, no significant difference existed between the monthly rates of terrestrial predator alarms (TPA) of Troop A and Troop B after controlling for group size (mean SE = 0.008 0.001 and 0.010 0.003 TPA/h per individual, respectively) (Z = 0.202, P = 0.843, nA = 12, nB = 12). However, while no correlation resulted between the monthly TPA rates of Troop A and the females monthly F levels ( rs = -0.01, P = 0.983), there was a st rong, positive correlation between the monthly TPA rates of Troop B and the females monthly F levels ( rs = 0.74, P = 0.022). Further, anecdotal evidence suggests that four of seven Troop A females produced their highest cortisol values 1 to 2 days following an encounter with a predator. The alpha female, Gina, produced her highest F level (84 times baseline) two days subsequent to a Harpy Eagle attack; J anes highest F level (9 times baseline) resulted two days following an (out of view) encounter with a terrestrial predator in the bamboo; and Carols and Kates highest F levels (13 and 26 times baseline, respectively) occurred the day af ter a close encounter with a jaguar. Troop Size, Stress, and Seasonality Troop A female cortisol levels were highest (mean SE = 71.60 13.46 ng/g) during the transition season and lowest (mean SE = 44.96 5.60 ng/g) during the dry season, while Troop B female cortisol levels were highest (mean SE = 59.73 48.00 ng/g) during the transition season and lowest (mean SE = 21.63 8.50 ng/g)
94 during the wet season. There were no statistical differences in seasonal intra-group F concentrations (Figure 3-4). Troop A females had a higher mean F concentration than Troop B females during the fruiting (mean SE = 60. 92 10.77 and 30.73 5.43 ng/g, respectively), wet (mean SE = 46.16 9.50 and 21.63 8.50 ng/g, respectively) and transition (mean SE = 71.60 13.46 and 59.73 48.00 ng/g, respecti vely) seasons, but Troop B females had a higher mean F concentration than Troop A females during the dry season (mean SE = 57.44 12.33 and 44.96 5.60 ng/g, respec tively). There were no statistical differences in seasonal inter-gr oup F concentrations (Figure 3-4). Troop Size, Stress, and Social Rank As described in Chapt er 2, the females in Troop A demonstrated a linear (although weak) dominance hierarchy. While the al pha female, Gina, had one of the lowest baseline cortisol values, and the most subor dinate female, Tinkerbell, had the highest baseline value, there was no linear correla tion between Troop As female baseline F concentrations and social rank ( rs = 0.14, P = 0.7599). In fact, with the exception of Tinkerbell, graphical representation demonstrates a bel l-shaped curve. A dominance hierarchy for the Troop B females could not be derived (Chapter 2), and thus I was unable to analyze the relationship between their social ranks and cortisol levels (Figure 3-5). Troop A females were involved in a signifi cantly higher rate of intragroup agonism, both as actors and recipients, than were the females in Troop B (Chapter 2). For neither troop, however, did the monthly rate s of female agonism given or received correlate with average monthly female cortisol values (Troop A: rs receive = 0.06, P = 0.863; rs give= 0.03, P = 0.914; Troop B: rs receive = 0.18, P = 0.638; rs give = 0.05,
95 P = 0.8994). For Troop A, however, the se asonal rate of agonism directed towards females displayed a strongly positive (althou gh not significant) relationship to female seasonal cortisol concentrations ( rs = 0.80, P = 0.20) (Figure 3-6). Discussion The results of my study indicate that over all, the females in the large group (Troop A) were more stressed (i.e., had higher cort isol levels) than the females in the small group (Troop B). Below I discuss the outco me of each hypothesis considering each variable of study and how it relates to the physiological effects of disparate group size. Group Size Affects the Level of Stress Females Incur Due to Predation Risk Support for this hypothesis varies according to the source of predation risk and verifies that (1) alarm calls are effective as ex trinsic indicators of in trinsic stress (Boinski et al. 1999) and (2) individuals perceive their ri sk from aerial versus terrestrial predators differently according to group size, as discussed in Chapter 2. In support of Cowlishaws (1997a) study t hat found that female desert baboons in small groups perceived the highest risk of predation, female C. apella in the small group were more stressed by the threat of terr estrial predators. In terms of aerial predation, however, it was the females in the large group that were most stressed by their perceived risk. It is of interest to note that while female cortisol concentrations in Troop A strongly correlated with both the physica l presence of a Harpy Eagle (actual predation risk) and the monthly rate of aerial alarms (perceiv ed predation risk), there was no correlation between Harpy Eagle presence and aerial alarm rates ( rs = -0.04, P = 0.904). This is likely because a Harpy Eagle attack was oft en not accompanied by an outburst of alarm calls; instead, the monkeys typical antipredator response to a Harpy Eagle is to suddenly drop lower in the canopy, often to the ground, and remain still and silent.
96 The distinction between actual and perceived predation risk directly relates to a distinction between shortand long-term stress. An encount er with a predator induces an acute stress response whereby energy is dive rted to the physiological functions most attuned to escape and survival. This adaptive stress response can become chronic (and thus maladaptive), howev er, when an individual percei ves her overall risk of predation to be high, even in the absence of a predator. If perceived risk accurately predicts the actual risk of predation, even if high, then the risk w ould be predictable and the stress response would likely be mediated by the down-regulation of glucocorticoid receptors in the brain (i.e., producing low co rtisol levels) (Sapolsky et al. 2000), as was found to occur in Beldings ground squirrels ( Spermophilus beldingi ) (Mateo 2007). The fact that Troop A females consistently had higher cortisol levels than their Troop B counterparts, and that their co rtisol metabolite concentrations were highly associated with monthly rates of predator alarms, ev en during a year of reduced Harpy Eagle activity, indicates that pr edation in RV is both strong and unpredictable enough to warrant a continued long-te rm stress response. Group Size Affects the Level of Str ess Females Incur During Food Shor tages This hypothesis is supported. It was the transition season, however, and not the dry season that was the most stressful time of year for both troops, and the females in the larger Troop A were significantly more stressed than the Troop B females during this time. I overlooked the importance of ( un)predictability on the impact of ecological stressors and originally ex pected the dry season to be the most stressful season for females due to a lack of fleshy fruits. The unpredictable nature of the transition seasons food supply, however, evoked more of a stress response than did a predictable lack of fleshy fruits, es pecially for the Troop A females.
97 Further, both troops average daily travel rate was highest during the transition season (Chapter 2), and, as tr aveling is a noisy and conspicuous behavior, their rates of aerial and terrestrial predator alarms were also the highest during the transition season (Chapter 2). Thus, a cascade of stress was created during the transition season, not only from the variable and unpredictable food supply, but also from the subsequent need to travel further in search of adequate nutrition and the associated increase in perceived predation risk. This supports the chronic stress hy pothesis (Boonstra et al. 1998) that predicts synergistic effects of food and predator pr essures (Krebs et al. 1995) and indicates that large-group female brown capuchins in RV are more susceptible to chronic stress. Further validating the impor tance of predictability in coping with stress is the finding that Troop A was actually the least stressed during the dry season. Posthoc consideration reveals a likely explanation for this unexpected result: The seasonal decrease in fleshy fruit availability was met by the troops increased use of the expansive bamboo patch centrally located within their home range. In fact, Troop A spent close to 60% of their time in the bam boo during the dry season, nearly double that of other seasons (Chapter 2). This bamboo patch not only provided consistently abundant (albeit lower quality), dispersed f ood resources (e.g., bamboo shoots and insect larvae) complements of the prior rainy season, but also provided the most protection from Harpy Eagles (a major st ressor to Troop A females). Another unexpected finding regarding the dry season is that females in Troop B produced higher cortisol concentrations than the females in Troop A. This uncharacteristically higher stress, however, does not appear to be explained by any of the social or ecological
98 variables studied, but instead was a likely r epercussion of the sudden loss of two of the troops females during this time. Thus, fema le cortisol concentrations for Troop B during the dry season were actually only representative of the single remaining female in the months following the disappearanc e of her two female conspecifics. The largest differential in cortisol conc entrations between the females in Troops A and B occurred during the wet and fruiting seasons. Troop B females were the least stressed during the wet season, the time of year that corresponds to the peak production of M. maripa fruits (Boinski et al., in prep). Fruiting M. maripa trees provided a large quantity of food in small, dense clusters and t herefore reduced the need to travel and forage. Additionally, being a small group, all Troop B individuals were typically observed to feed simultaneously without a s ubstantial increase in agonism. On the other hand, the larger Troop A experienced more stress during the wet season, even with a substantial ( M maripa -related) decrease in daily travel rate (Chapter 2). As opposed to Troop B, the females in Troop A experienced an intensified dominance hierarchy and a subsequent increase in tensi on while feeding in the small clusters of this high-quality fruit. Females were the recipients of significantly more interand intra sexual agonism from adult c onspecifics while feeding on M maripa (Chapter 2), and the seasonal rate of agonism directed to wards females was positively (although not significantly) related to female cortisol c oncentrations. I suspect that the increase in stress demonstrated by the Troop A females during the dry season was largely due to the anticipation of agonism, even more so than the actual aggression itself, that can be expected during intensified intragroup competition.
99 Rank-Related Stress is a Factor of Group Size This hypot hesis is supported, primarily due to the fact that a dominance hierarchy was determinable only among the Troop A female s. The alpha female had one of the lowest basal cortisol levels while the mo st subordinate female had the highest basal cortisol level, and it is of interest to note that no females basal cortisol concentration differed significantly from that of the female directly above or below her in the hierarchy. However, no linear relationship was found between female rank and basal cortisol level, and the largely bell-shaped relationship may act ually indicate a cost of being a middleranked female in such a large troop. Estab lishing linearity to the hierarchy, however, was most difficult for the three middle-ranking females and seemed rather moot, and so I suspect that the bell-shaped curve is an artifa ct of individual va riation rather than indicative of rank-related stre ss. Chapman et al. ( 2006) predicted that high variance in cortisol levels indicates stress and is in some way a function of dominance. My study provides support for this assumption, as t he most subordinate female (Tinkerbell) not only produced the highest basal cortisol level but also the most variance in her cortisol measures. The females in the large troop were clearly more susceptible to the effects of a dominance hierarchy than were the females in the small troop; thes e effects, however, were most evident when the food supply was unpredictably available and/or distributed in small clumps. Otherwise, rank-related interactions did not appear to be a major source of stress for females, as was shown in rodents (Tam ashiro et al. 2005) and other female-bonded species such as baboons ( Papio spp. ) and macaques (Macaca spp.) (Sapolsky 2005). The (weakly) linear hierarchy among Troop A females was rarely enforced, with little overt aggr ession between females.
100 Conclusion Optimal group size for a species is theor etically assum ed to be of intermediate size (Silk 2007a). So how does one explai n the long-term maintenance of suboptimal (and disparate) group size, as demonstrated by these two troops of brown capuchins? Just as females in female-philopatric soci eties tend to remain in their (sometimes suboptimally-sized) natal troop because the benef its of staying outweigh the costs of migration, a group is not likely to bear the costs of adjusting size (e.g., re-establishing territories and dominance relationships) in order to better fit environmental challenges when the environment is so dynamic and the challenges so unpredictable (Pride 2005a). In this chapter, I demonstrated that group size and ecological context interact to affect an animals internal state, which may be a critical proximate factor linking these environmental pressures to individual fitness (Pride 2005a). In particular, cortisol data suggest: Females incur a cost-benefit tradeoff of disparate group size. A large group may provide a more stable system for coping with major socioecological fluctuations over the long-term, but incur increased stre ss in day-to-day life, as demonstrated by the high cortisol values of the Troop A females. A small group, such as Troop B, is less dynamic; individual group me mbers better endure the daily challenges of ecological variation, but are vulnerable to any major social upheaval or environmental shift. The sensitivity of sm all-group females to large-scale social disruption occurs in rodents (Haller et al. 1999) and was evident in my study through the increase in cortisol concent ration and the prolonged increase in the variance of cortisol of the only remaining female in Troop B in the months following the disappearance of her two female conspecifics. In a high risk, low energy environment su ch as RV, female brown capuchins in large groups are susceptible to chronic stress, particularly in response to predation risk and an unpredictable food suppl y. These females are assumed to be physiologically cha llenged (Chapman et al. 2006) and may face long-term fitness consequences. In the following chapte rs, I investigate in further detail the nature of the fine-scaled social relationships of females in Troop A to determine
101 the social, ecological and reproductive fa ctors most closely associated with their stress levels.
102 Table 3-1. Baseline values of individual female fecal cortisol concentrations. Each females baseline value is the calcul ated mean after removing the highest and lowest 5% of values, and was used to analyze the effect of state (long-term) variables, such as troop size, domi nance rank and seasonality, on cortisol concentrations. The baseline cortisol concentration (ng/g) of Troop A was (non-statistically significant) higher than that of Troop B. Troop Female n F baseline range (ng/g) F baseline value (ng/g SE) A Kate 34 3.3 252.4 28.75 7.46 Gina 29 7.2 200.4 37.21 8.73 Carol 37 7.3 298.0 49.08 10.15 Little Horns 40 9.9 261.5 57.40 9.28 Banana 38 4.3 271.2 61.72 10.30 Jane 27 10.5 286.1 68.04 14.15 Tinkerbell 32 7.2 445.1 75.41 20.88 TROOP A 237 3.3 445.1 53.86 4.58 B Mrs. Slocum 10 5.4 79.9 25.34 7.05 Luna 5 29.0 86.2 44.38 .61 Miss Brahms 19 6.1 155.7 48.58 10.20 TROOP B 34 5.4 155.7 41.13 6.40
103 Figure 3-1. Troop-level fecal cortisol (ng/g SE) profiles of Group A and Group B females, based on average monthly baseline values. Troop A : n=7 (January to September), n=6 (October to December); Troop B : n =3 (January to July), n=1 (August to December). Values are not included for Troop B in the months of January, June, July and December due to insufficient data. Although Troop A females tended to have cortisol concentrations that were, on average, higher than Troop B female cortisol levels, the two troops demonstrated similar patterns of monthly variation in their females baseline cortisol profiles. Figure 3-2. Morning versus afternoon cortisol levels (ng/g SE) for Troop A and Troop B females. Although the females in Troop A demonstrated higher morning and afternoon averages than the Troop B fema les, no statistical interor intratroop difference existed in the females morning and afternoon cortisol levels. 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 JanFebMarAprMayJunJulAugSeptOctNovDec Troop A Troop B 0 10 20 30 40 50 60 70 morningafternoon Troop A Troop BX cortisol concentration ( n g / g) X cortisol (ng/g)
104 Figure 3-3. Monthly comparison of the rate of aerial alarms (AA/h) produced by Troop A, observed number of Harpy Eagle attacks and/or attendance to the troop within 50 m, and mean cortisol levels of Troop A females. A strong, positive correlation existed between the monthly rates of AAs and the females monthly cortisol levels ( rs = 0.74, P = 0.009). An even stronger correlation existed between the monthly presence of Harpy Eagles and the females cortisol levels ( rs = 0.80, P = 0.002). 0 1 2 3 4 5 6 7 JanFebMarAprMayJunJulAugSepOctNovDec Cortisol/100 (ng/g) AA rate (alarm/h) # Harpy Eagle visits
105 Figure 3-4. Seasonal comparison of average baseline cortis ol levels (ng/g SE) for Troop A and B females. The number of females present in each troop during each season is indicated in parentheses. Troop A female cortisol levels were the highest during the transition seas on and lowest during the dry season, while Troop B female cortisol levels we re highest during the transition season and lowest during the wet season. Troop A females had a higher mean cortisol concentration than Troop B female s during all seasons except the dry season, a time confounded by the disapp earance of two of the three Troop B females. The largest intergroup differential in female cortisol concentrations occurred in the fruiting and wet seasons when fruits were distributed in small clumps. 0 20 40 60 80 100 120 Fruiting Wet Dry Transition Troop A Troop BCortisol (ng/g) (7) (7) (7) (6) (3) (3) (1) (1) (Feb-Apr) (May-Jul) (Aug-Oct) (Nov-Jan)
106 Figure 3-5. Mean baseline cort isol concentrations (ng/g SD) of individual adult females in Troop A by their respective dominance rank (in parentheses). Mean baseline cortisol concentrations of Troop B females are also shown for comparison, although a dominance hier archy could not be established among them. The females in Troop A demons trated a (weakly) linear dominance hierarchy, although their respective basel ine cortisol levels did not correlate with rank. Still, however, the alpha fe male had one of the lowest baseline values while the most subordinate fema le had the highest baseline value with the most variance. Troop A Troop B Gina (1) Banana (2) Jane (3) LH (4) Carol (5) Kate (6) Tink (7) Luna Slocum Brahms-100 -50 0 50 100 150 200 250 Cortisol (ng/g)
107 Figure 3-6. Seasonal cortisol concentrati ons (ng/g SE) of Troop A females and the rate at which females were the recipien ts of social agonism from conspecifics (bouts/h). The seasonal rate of agonism directed towards females was strongly and positively (although not signific antly) related to female cortisol concentrations (rs = 0.80, P = 0.20). 0.002 0.004 0.006 0.008 0.01 0.012 507090110X female cortisol concentration (ng/g) X rate of seasonal agonism directed towards females (bouts/h) Dry Fruiting Transition Wet
108 CHAPTER 4 FEMALE-BONDED PRIMATES? GROOMI NG AND PROXIMITY PATTERNS OF FEMALE BROWN CAPUCHINS IN SURINAME Introduction Despite the abundant use of the term social bond (or one of its many variants) in published articles and the growing literatur e regarding the func tional, adaptive, and theoretical consequences of bonds, there re mains a fundamental s ource of confusion that precludes systematic comparison of related research: What is a bond? The term bond (a.k.a. social bond, relationship, friendship) is loosely defined, often not operationalized, and used to characterize a variety of social situations (Figure 4-1). For example, bond can be synonymous with pai r bond or cross bond between the sexes, in which the term is typically used to describe non-random sexual behavior. For same sex relationships, however, the usage of bond becomes more ambiguous, and many authors do not clearly define or consistently apply their use of the term (Figure 4-1). Female-bonded, for example, could indicate ei ther the presence of female philopatry, or differentiated, affiliative relationships between females, which may or may not be kin-based. Using Affiliative Behavior as a Proxy Measure of Social Bonds When social bond is used to describe a highly affiliative relationship between two individuals, relationship strength is typically measured as groomi ng investment and/or time spent in proximity. Females with t he strongest grooming relationships have the strongest and most enduring social bonds (Silk 2007b), and non -random proximity patterns are suggested to indicate indirect affiliation between individuals (Fragaszy et al. 2004). As such, female-bonded primates are expected to groom and spend more time in close proximity with other females than with males.
109 But are grooming and proximity maintenance equivalent measures of relationship quality, or does use of varying indicators further confound the ambiguity of social bonds? Grooming relationships, by default, include the costs and benefits associated with proximity maintenance; the reverse, however, cannot be said In fact, Henzi et al. (2003) reported that female baboons did not preferentially groom those females that most frequently maintained nearest neighbor pr oximity. Is a dyad that demonstrates higher than average time spent in proximit y but does not groom considered socially bonded? Whether or not grooming and proximity are even truly indicative of a bond is also debated. Barrett and Henzi (2002) sugges ted that both grooming and proximity patterns are labile and therefore not indi cative of a long-term bond but rather are currencies to be traded as a short-term strat egy. This brings to light another issue of ambiguity concerning social bo nds: Just how stable or lasting must a relationship be before the participants are truly bonded? Are long-term bonds (friendships) and short-term bonds (business partnerships) r eally so different? There must be some adaptive advantage to the benefits of non-random dyadic affiliation, no matter how fleeting, or else the parti cipants would not invest the time and energy into such behavior. Whether a social grou ps relationships are shortor long-lasting is intriguing in itself, and understanding the nature of and variation in bonding patterns, especially variation within-species, will help us to understand the socioecological factors underlying and influencing sociality. A Female-Bonded Primate? The distribution of females and female rela tionships is widely understood to be at the core of primate social groups (Emlen and Oring 1977; Mi tchell et al. 1991; van Hooff
110 and van Schaik 1994); the structure of female social relationships, however, is less understood. While female relationships vary according to local circumstance and in response to the current tradeoffs of the cost s and benefits of dyadic association (Barrett et al. 2002), female social bonds are typically presented as a dichotomous, species level characteristic. Capuchin monkeys (Cebus spp. ) present a propitious system to test social evolution models derived largely from Old Worl d monkeys (Parr et al. 1997). Capuchins are unique among platyrrhines in that female philopatry is clearly prevalent throughout the genus, a feature common to catarrhines Although female philopatry is not a necessary condition for the development of social bonds between females (Byrne et al. 1990; van Schaik 1989), a female philopatri c group provides a study system conducive to long-term observation of females with inve sted relationships and a history of repeated interaction. Not only are all Cebus species female-philopatric, but all are also categorized as strongly female-bonded (Janson and Boinski 1992; Fedigan 1993; OBrien 1993; Perry 1996; Di Bitetti 1997; Manson et al. 1999). For example, among C. capucinus females spent more time in proximity to other adult females than adult males, and females groomed other females twice as often as th ey groomed males (Perry 1996). Similarly, Di Bitetti (1997) concluded that female C. apella in Argentina established stronger grooming bonds between each ot her than did males. Do such findings, however, substantiate classification fo r the entire species, or could such social behavior be siteor time-specific? Recent finding from Brazil introduced a C. apella population that lacked female grooming bonds (Izar 2004), and preliminary observations of female
111 C. apella in Suriname suggested both temporal and individual variation in intrasexual relationships (Boinski, unpublished data). In the previous chapters, I examined troop-level socioecological and endocrinological variables related to the cost s and benefits of socialit y for female brown capuchins living in groups of disparate size In this chapter, I focus on fine-scaled within-group variation in female social behavior and quantitatively determine the structure of and temporal shifts within their relationships. For pur poses of my study, I defined a social bond as the outcome of si milar affiliative strategies between two individuals in a socially cohesive group, and I operationally defined a female bond as one in which a females primary adult so cial partner, measur ed through grooming and proximity, is another female. Hypotheses Social Bonds among Female Brown Capuchins are Variable Over Time Social relationships are identifi ed through the frequency and patterning of interactions over time, and a long-held premis e underlying primate sociality is that in stable, coherent groups, strongly different iated and enduring relationships between females will always be detectable (Henzi et al. 2009). Silk et al. (2006b) measured bond stability as the number of consecutive years in which a female was among another females top three affiliative partners; they determined that, among savannah baboons ( Papio cynocephalus ), females demonstrated considerable annual stability in their social bonds with some dyads (typically of related females or females of similar age and rank) maintaining close bonds for at least 3 years, and some for mo re than 5 years. While they acknowledged the presence of more ephemeral social bonds, Silk and her colleagues focused on the longer-lasting relationshi ps that they characterized as strong
112 family ties and friendships. How much variation was lost, however, when examining dyadic behavior in annual groupings? The expression of affiliative behavior vari es according to ecological conditions (Schino and Aureli 2008), and Barrett et al. (2002) argued that female relationships are more variable than stable, with females expl oiting temporary, short-term social relationships. Female chacma baboons ( Papio cynocephalus ursinus) residing in a seasonal environment were more business partners than friends, demonstrating monthly and yearly inconsistencies in t heir preferred grooming partner and using grooming behavior as a commodity to be bartered for their short-term advantage (Barrett and Henzi 2002). Preliminary observations of the brown capuchins in RV indicated that female C. apella social relationships are more variabl e than is expected for a female-bonded species. In the years prior to my study, adult females were initially observed to focus their grooming on adult males and only rarely groom other females. A short time later, however, patterns of female-female affiliation shifted, becoming more typical of what is expected of C. apella females (Boinski, unpublished data). Thus, both temporal and individual variation in t he intensity of female C. apella relationships exists at RV. I predicted that females do not consistently demonstrate stronger gr ooming and proximity relationships with other females than with adult males. Dyadic Female Relationships Vary in Response to Ecological Pressures Given the premise that individuals form and maintain relationships in response to potential benefits, patterns of social interact ion should reflect the desirability of partners and the benefits provided by the relations hip (Schino and Aureli 2008). In an environment characterized by social and ecolog ical change, I contend that investment in
113 relationships should follow a similar pattern as circumstances change, so do the benefits of particular dyadic interactions. Henzi et al. (2009) provided empirical evidence that female chacma baboons ( Papio hamadryas ursinus ) displayed seasonal changes in social relationships in response to ecological conditions. At times, most notably during the food-scarce season, the fe males formed short-term companionships with various partners; during the food-abundant season, however, females displayed no particular affiliative partner connections and were best viewed as simply gregarious. Social bonds are suggested to be instrum ental in improving access to resources by creating tolerance and securing agonist ic support (Henzi and Barrett 1999), and tolerance by dominant individuals may significantly increase feeding success of subordinates, especially in small patches (J anson 1990a). In fact, a meta-analysis of dominance hierarchies, grooming patterns, and ecological conditions determined that steeper dominance hierarchies were associ ated with more (non-reciprocal) grooming being directed up the hierarchy (Schino and Aureli 2008). Therefore, social bonds can be viewed as a proactive strategy to counteract ecological variation. As was shown in the previous chapte r, the transition season (November to January) was the most stressful time of year for female C. apella in RV, presumably due to the unpredictable nature of the food supply; the wet season (May to July), however, was the time of year that Troop A females received the most agonism from troop-mates, presumably due to the peak fruiting of a food resource that is distributed in small, monopolizable clusters. Therefore, I pr edicted that females non-randomly increase affiliative behavior during the transition and wet seasons, and that the intensified grooming would be directed up the hierarchy more so t han in the other seasons.
114 Relationships among Females Strengthen w ith the Presence of an Infant Among baboons ( Papio cynocephalus ursinus ), infants are used as a commodity within a biological marketplace, such that non-mother females gr oom the mothers of young infants in exchange for access to the in fants (Henzi and Barrett 2002). Providing an adaptive explanation for this behavior, Strier (1999) suggested that strong female relationships are vital for primates as their slow-paced developmental patterns encourage cooperative infant rear ing that is exemplified by their carrying and nursing of infants other than their own (Fragaszy et al. 2004). Further, Silk et al. (2003) provided empirical evidence that infant survival wa s significantly associ ated with the level of social integration of post-partum females. Therefore, I predict ed that females with young infants (< 3 months old) would be more socially integrated than non-mothers, and that mothers would be more likely to groom and/or be groomed by other females in response to the physical presence of their infant. Methodology Site Description Raleighvallen Nature Preserve (RV), a 7812 km2 reserve consisting of primary tropical rain forest, is located within the 1.6 million ha Central Suriname Nature Reserve in Suriname, South America. The main study site measured approximately 2 km2 and was covered by an extensive trail system RVs flora and fauna are effectively undisturbed in the historical period, and an intact array of potential predators and competitors are present (Reichar t 1993). The forest in RV is ecologically distinct as compared to other areas where brown capuchi ns are commonly studied. The soils of the Guyanan Shield (including the Guianas and Suriname) are nutrient-poor and highly weathered as they are derived from Precambrian bedrock. Fleshy fruits are low in
115 abundance and available fruit resources typi cally occur in small patches (< 5 m diameter) (Boinski et al. 2002). Additional ly, much of RV is composed of expansive bamboo patches and dense liana forest that restricts visibility (Boinski et al. 2003). Study Animal The brown capuchin monkey ( Cebus apella ) is one of the four traditionally recognized species in the genus Cebus (Hill 1960). Capuchins li v e in multi-male, multifemale polygamous groups that are typically female philopatric. Adult males and females maintain separate dominance hierar chies; in general, males are dominant to females. C. apella social groups are typically compri sed of 12 to 27 individuals, with an average troop size of 17 members. Troop membership and relationships remain relatively stable, except for subadult male s who emigrate from their natal group and may subsequently transfer several times. Brown capuchins reside in a variety of habitats encompassing a large geographic range from Columbia to Argentina (Fragaszy et al. 2004). C. apella are omnivores; the bulk of t heir diet consists of fruit supplemented with seeds, vegetation, arth ropods, and vertebrate s (Terborgh 1983). The study troop (Troop A) was one of four brown capuchin groups commonly observed within the study site, and the primar y study subjects were the adult females and their relationships to the adult members of the group (Table 4-1). For most of the study, Troop A consisted of 27 individuals, se ven of which were adult females. During the study, one pregnant fema le disappeared; her dis appearance coincided with her predicted due date, and thus it is possible her (presumed) death was related to pregnancy or labor-related complications Relatedness among troop members is unknown; however, genetic analysis of the brow n capuchin population in RV is currently underway. The study troop had been habituated to human observers since 1998, and
116 all individuals were recognize d based on body size, color patterns, and other identifying characteristics such as scars, moles, and ear shape. Throughout the study period, inter-observer reliability tests were conducted monthly to ensure consistent and accurate identification of all troop members. Data Collection From January to December 2006, I and/or field assistants followed the study troop for up to 10 h each day (beginning at sunrise). Troop A was followed, on average, 28 days per month, comprising a total of 2284 h of observation (not incl uding out of view time). The following behavio ral data were collected: (1) Individual instantaneous scans In or der to determine each females primary proximity partners and social network shifts over time, I collected instantaneous scans each time a female was in view and her neares t neighbors could be reliably identified. Each scan effectively provided a snap-shot of that moment in time, including the females location in the troop, her behavior, and the identity and behavior of all visible neighbors. Scans were not taken on females involved in direct social behavior such as grooming or aggression; thus, the scans repr esent an indirect measure of affiliation based on close proximity during travel, feeding, resting, etc. Furt hermore, a female was never scanned while she was in estrus, as that would likely bias the data towards male neighbors. Peripheral females were scanned as frequently as were the more central females. A total of 787 scans were collected on Troop A females throughout the study. (2) Ad libitum notes. C. apella social behavior (agonistic, grooming, and sexual) accounts for <10% of their daily activity budget (Zhang 1995a). Therefore, ad libitum sampling (Altmann 1974) provided a more efficient and complete data collection technique for capturing social interactions than did a more structured sampling regime.
117 Social behavior, such as grooming, play and agonism, as well as other opportunistic occurrences such as predator interacti ons, intergroup encounters, and reproductive behavior were noted ad lib. To reduce observa tion bias, I moved constantly through the groups to locate non-visible individuals and reco rd their activities. Due to opportunistic sampling, as well as the reduced visibility wit hin the Raleighvallen forest, the recorded frequencies of social behavior are likely unde restimates of the true frequencies. Because of the large number of hours over which behavioral interactions were recorded, however, I am confident that the data provide strong indication of actual trends. Behavioral Definitions A females social network was defined as the sub-group of conspecifics most frequently observed within a 4 m proxim ity. A female bond was defined to occur when a females primary adult social partner (in terms of grooming and/or pr oximity patterns) is another female. An adult dyad was considered to have a strong bond (based on grooming or proximity) if t heir dyadic grooming duration or proximity score (see below) fell within the top 10% for t he group (Silk et al. 2006a). Grooming bout duration and frequency was recorded for all observed occurrences. A grooming bout involved one individual gr ooming another, and concluded when the grooming ended or was interr upted by another behavior for more than 5 seconds. A new grooming bout began with a new dyad, or whenever the grooming continued or changed direction within the same dyad. Only grooming bouts involving at least one female (as actor or recipient), and where both participants could be i dentified (at least to the age and sex class) were included for analysis. All grooming bouts were counted as dyads; thus, if two monkeys were grooming the same individual concurrently, the
118 episode was scored as two separate actorrecipient grooming bouts. Grooming bouts involving females in estrus (as actors or receivers) were not included in analysis. Agonism includes all instances of aggressive, displacement, and submissive behaviors. Aggressive behaviors include c hase, hit, bite, and threat display. Displacement involved one individual passively supplanted by another, usually (but not exclusively) at a food source; if an indivi dual aggressively displa ced another individual, then the bout was scored as aggression and not displacement. Submissive behaviors included any instance in whic h an individual retreated and/or reacted with submissive posture or vocalizations without being aggressed towards or displaced. Only agonistic bouts involving at least one female (as acto r or recipient), and where both participants could be identified (at least to the age and sex class) were included for analysis. All agonistic events were counted as dyads. If the agonism involved a coalition against a third party, only the initial inst igator and victim were scored. Agonistic events involving females in estrus (as actors or rece ivers) were not included in analysis. Data Analysis Dominance ranks Female ranks were determined by t he outcomes of decided female-female agonistic episodes. All agonistic bouts were categorized as either decided or undecided. I considered an agonistic bout as decided if: (1) ident ities of participants were known; and (2) one individual displayed only aggressive signals whereas the other only displayed submissive behavior, or (3) one individual displayed submissive signals, whereas the other individual di splayed no agonistic behavior. All other disputes were considered to be undecided (Silk et al. 2006b). Dyadic agonistic bouts between
119 females were constructed into a dominance ma trix to determine the hierarchical nature of female relationships (Ferreira et al. 2006). Proximity scores To determine females primary proximity partners and social network shifts over time, I collated the scan data into monthly patte rns of association. A females social network included thos e conspecifics most fre quently within a 4 m radius (a post-hoc, empirically-based boundary). Mo nthly proximity scores for eac h female were calculated (1) for each adult dyad, and (2) by the age/sex class categories of neighbors: adult males (MPS), adult females (TFPS), and all neighbors within 4 m ( adult males, adult females, and immatures) (T PS). Additionally, as 4 of the 7 females had dependent infants in the initial months of the study, I was able to invest igate the social effect that young infants have on their mothers. I calc ulated proximity scores separately for the mothers based on the physical presence of thei r young infants (i.e., while they were and were not carrying their infants) and also differentiated female neighbors that were carrying an infant (FwiPS) from female neigh bors that did not have an infant (FPS). Following Perry (1996), proximity scores were calculated as follows: For each adult dyad (and age/sex class), I tallied (by mont h) the number of scans that a particular adult male or female individual was within 0 to 1 m, 1 to 2 m, 2 to 3 m, and 3 to 4 m of each female. Neighbor presence was not mut ually exclusive, and therefore multiple neighbors could be included in one female sc an. Because close neighbors were less likely by chance to be near a female than ar e more distant neighbors, I multiplied the four scores by weighting factors, such t hat the closer proximity categories were weighted more heavily. Weighting factors were determined using concentric spheres, with the focal female in the middle, such that the innermost sphere has radius one, the
120 second sphere has radius two, the third s phere has radius three, and the outer sphere has radius four. The weighti ng value for the 0 to 1 m category is the reciprocal of the volume of a sphere of radius one; the weight ing value for the 1 to 2 m category is the reciprocal of the differenc e between the volume of a sp here of radius two and the volume of a sphere of radius one, etc. For comparative purposes these four scores were grouped into a single proximity score. Thus, the equation for calculating the proximity score is: Proximity score = 1,000(0.239a + 0.034b + 0.0126c + 0.0065d), where a is the proportion of a females to tal monthly scans in which a neighbor was within 0 to 1 m of the scanned female, b is the proportion of scans in which a neighbor was within 1 to 2 m of the scanned female, and so on. When calculat ing proportions for female-female dyads, the total number of scans for each of the females in the dyad was summed. The sum of the four weighted values is multiplie d by the arbitrary value of 1,000 to make the proximity score more rea dable. The higher the proximity score, the more time the female spent in close proximity to that individual (or age/sex class). Grooming Data were only inc luded in this analysis if the identities (for adults) or age class (for immatures) of the par ticipants were known. All grooming bouts were counted as dyads; thus, if two monkeys were grooming the same individual concurrently, the episode was scored as two actor-recipient grooming bouts within the same grooming session. In order to maintain consist ency with the proximity data, grooming bouts involving females in estrus (as actors or recipients) were not included in analysis. Statistical analysis Both parametric and non-parametric analyse s were used. Non-parametric analyses were used when data failed to meet the assumption of normality.
121 Homoscedasticity was tested using Levenes Test for Equality of Variances; if Levenes test indicated heterogeneous variances, I used square root transformations. I used the Mann-Whitney U test when comparing tw o independent samples and the Wilcoxin sign rank tests for comparing two related samples. When comparing more than two samples, I used either one-way ANOVA in conjunction with Tukeys post hoc tests or, when data remained non-normal even afte r logarithmic transformation, the Kruskal Wallis one-way ANOVA in conjunction with Mann-Whitney U tests as the non parametric alternative to posthoc tests. To investigate whether an observed frequency distribution fit an expected one, I used the Chi-square goodness of fit test. Correlations between parameters we re evaluated using Spearmans rs (Excels Analyse-It statistical package). Analyse s were performed using SPSS 11.0; the significance level for all tests (two-tailed) was set at 0.05. Mean values are presented with standard error values unless otherwise noted. Results Dominance Hierarchy Only 27 decided agonistic bouts between females were observed (0.01 bouts/h), of which 63% involved displacement, 30% were aggressive, and 7% involved submissive behavior only. Due to low val ues the dominance matrix of female-female agonistic frequency was not possible to test statistically. T herefore, dominance relationships among females were not so lely determined by the frequencies of dominance interactions. Among the 21 dom inance relationships, 13 were decided based on agonistic frequency. Zero-ties acc ounted for eight dyadic relationships, of which four were decided by applying the transit ive property, such that a > b, b > c, therefore a > c (Izar 2004). The dominance relationship between Banana and Jane,
122 however, was decided based not on the in teractions between them, but on the observation that Banana aggressed towards the dominant female without receiving retaliation. Jane was never observed to ch allenge the dominant fe male, and therefore Banana received a higher ranking than did J ane. The remaining unknown relationships (especially between Kate and Tinkerbell) ste mmed primarily from Tinkerbells lack of interaction with other females. Tinker bell was presumed to be the youngest of the groups adult females, and remained periphe ral to the troop throughout the study (except during estrus). In fact, she scored a TFPS near 0.0 and was the female least involved in female-female grooming (see below). Therefore, I ranked Tinkerbell as the least dominant female (Table 4-2). Female Coalitions Females were observed to participate in 46 coalitions, 40 of which were dyadic and the remaining six were triadic. Fifty-fi ve percent of dyadic coalitions involved a male-female pair, 35% involved a fema le-female pair, and 10% involved a female-juvenile pair. The most frequent ( 26%) coalition partners were the alpha male and the alpha female (Gina). The second mo st frequent (21%) coalition partners were the two highest ranking females (Gina and Banana) The six triadic coalitions involved various groupings of Gina, Banana, the alpha male, and one of the subordinate males. Coalitions were formed to threaten conspecif ics, mob potential dange rs (e.g., snakes), and threaten human observers. Females were never observed to form coalitions against another female; three of the female-f emale coalitions were directed at an adult male.
123 Troop Patterns of Affiliation Proximity Overall, adult female proxim ity scores with males (MPS) were significantly higher than female-female proximity scores (TFPS), despite a 3:7 male to female sex ratio ( Z = 2.301, P = 0.021). Female MPS scores increased with female rank, while there was no clear pattern between TFPS scores and female rank. The alpha female demonstrated a higher average TFPS score than MPS score while the remaining females had higher average MPS scores than TF PS scores. Statistical comparison of scores, however, yielded only two significant results: Little Horns and Tinkerbell had significantly higher MPS scores than TFPS scores ( ZLittle Horns = -2.395, P = 0.017; ZTinkerbell = -2.668, P = 0.008). On average, the two mo st dominant females had total proximity scores (TPS), including female, ma le and immature neighbors, almost three times higher than the subordi nate females (Table 4-3). Grooming I recorded 1574 dyadic grooming bouts ( with known direction and participants) between Troop A members (0.69 bouts/h), 1013 of which (64%) involved at least one female. On average, females spent 1.8% of their monthly activity budget on grooming (as actors and/or receivers). The amount of time females spent grooming adult troop-mates was in (approximate) positive asso ciation with female ra nk. The amount of time that females received grooming from ad ult members of the tr oop, however, did not appear associated with rank (Figure 4-2). Females groomed up and down the hierarchy at equal rates; there was no significant diffe rence in the monthly rate (min/h) that subordinate females groomed dominant fe males (mean SE = 0.030 0.008) and
124 dominant females groomed subordinate females (mean SE = 0.029 0.009) ( Z = -0.11, P = 0.909, n = 26) (Figure 4-3). Overall, females gave a significantly higher rate (min/h) of grooming than they received ( Z = -3.059, P = 0.002, n = 12). Females acted as the groomer in 817 bouts (81%) and as the recipient in 371 bouts (37%). Females devoted 1937.7 min to grooming conspecifics, with juveniles rece iving the largest percentage (36.1%) and females receiving the least (19.1%). Female s received a total of 755.1 min of grooming from conspecifics, the larges t percentage of which was received from females (49.1%). Not including infants, adult ma les groomed females the least (17.2%) (Table 4-4, Figure 4-4). Duration of adult grooming bouts significantly differed by sex ( F = 5.767, P = 0.003, df = 2). Females groomed males for a longer duration per bout (mean SE = 3.1 0.271 min) than fema les groomed females (mean SE = 2.1 0.165 min) ( P = 0.022) (Figure 4-5), and females gr oomed males longer per bout than males groomed females (mean SE = 2.1 0.323 min) ( P = 0.008). Only 30% of the adult grooming bouts were reciprocal. Fe male-male grooming dyads tended to be slightly more reciprocal than were fema le-female grooming dyads (26.1% and 24.7%, respectively). Individual Patterns of Affiliation Proximity Following Silk et al. (2006b), I cl assified the dyads that fell within the top 10% of all proximity scores as having strong bonds. Although the 10% subset may seem arbitrary, it was also a clear empirical br eak in the data: Str ong bonds resulted from dyadic proximity scores higher than 100; the nex t highest proximity score was 83. Of
125 the 21 possible female-female dyads, only one was represented in the strong bonds, whereas 3.5 female-female dyads fell within t he lowest 10% of proximity scores. By contrast, of the 21 possible female-male dyads four resulted in strong bonds, and none fell within the lowest 10% of scores (Appendix C). Those strongly bonded dyads included, in descending order: Gina and Boris (272.49); Gina and Banana (233.69); Banana and Gina (233.27); Banana and Boris (151.67); Jane and Boris (114.05); and Tinker bell and Boris (110.03). Thus, the only strong (proximity) bond observed between females (Gina and Banana) was also an equitable one. The remaining strong bonds were represented by the three highest-ranking females and the alpha male, with one exception. The strong bond between Tinkerbell and Boris wa s surprising, given her low rank and overall lack of gregariousness (Table 4-5). Tinkerbells hi gh proximity score with Boris actually resulted from three uncharacteristic occurre nces (further confounded by her overall lack of proximity data) and should not be taken at face-value. Monthly measures of her proximity to the alpha male varied drastica lly, with a mean monthly value of 11.0 and standard deviation of 17.31; thus the mean is an unreliable predictor of Tinkerbells true proximity-based relations hip with Boris. Grooming Following Silk et al. (2006b), I cl assified th e dyads that fell within the top 10% of total grooming duration (min) as havi ng strong bonds. Of the 21 possible female-female dyads, only one was (reciprocally) represented in the strong bonds. Four female-female dyads were never observ ed to groom, and in four additional female-female dyads, one partner was never observed to groom. Of the 21 possible female-male dyads, five strong grooming bonds occurred: one that was reciprocal within
126 the dyad, and four that involved unidirectional grooming of the alpha male by females. Five female-male dyads were never observed to groom, interestingly three of which involved the three highest-ranking female s (who were strongly bonded to the alpha male through grooming and proximity) and the same subordinate male. Tinkerbell was the only female that did not groom the alpha male, Boris, and she was the only female that he was not observed to groom. In five additional female-male dyads, one partner was never observed to groom (Appendix D). The strongly bonded dyads included, in desc ending order (groomer and recipient): Gina and Boris (86.5 min); Gina and Banana (83.8 min); Banana and Boris (82.5 min); Kate and Boris (51.2 min); Banana and Gina (49.8 min); Little Horns and Darwin (40.6 min); Jane and Boris (36. 3 min); and Darwin and Little Horns (30.2 min). The only strong (grooming) bond observed between females (Gina and Banana) was also an equitable one; the only other equitable st rong grooming bond was that between Little Horns and Darwin. All other strong groom ing bonds involved a female (the three highest-ranking females and a low-ranking female with a young infant) grooming the alpha male, and these grooming relationships were considerably inequitable (Table 4-6). Grooming versus proximity Dyadic total proxim ity scores and grooming durations were modestly correlated ( rs = 0.41, P = 0.0010, n = 63) (Figure 4-6), a pattern that held true when considering female-female ( rs = 0.41, P = 0.0068, n = 42) and female-male ( rs = 0.48, P = 0.0266, n = 21) dyads separately. When only consid ering those dyads that were strongly bonded through both grooming and proximity (Gina/Boris, Gina/Banana, Banana/Boris,
127 Jane/Boris), there was a str ong association between total pr oximity score and grooming duration ( rs = 0.90, P = 0.0374, n = 5). Monthly Variation in Affiliation Proximity There was no signific ant monthl y variation in overall TFPS ( H = 4.453, P = 0.879, df = 9) or MPS (H = 9.386, P = 0.402, df = 9). Sample sizes were too small to test statistically for monthly variation in each fe males dyadic proximity scores. Patterns emerged, however, to suggest t hat the females primary adul t proximity partners did not remain consistent over time (Table 4-7). Strong bonds, i.e., the top 10% of monthly dyadic proximity scores (Silk et al. 2006b), were exhibited within 21 dyads: 6 and 15 pairs; 21 dyads failed to exhibit a single strong (mont hly) proximity bond during the length of the study. The most consistent and reciprocal of the strong proximity bonds was between Gina and Banana, the two highest ranking females. The remaining majority of strong bonds were demonstrated between a female and a ma le. Gina and Banana had the most (5) monthly strong bonds with Bori s, although those bonds were inconsistent and without discernable pattern. Gina, B anana, Jane, Little Horns, and Carol each had two months of strong bonds with Darwin, the youngest of the adult males. Again, the bonds were expressed rather inconsistently but it is interesting to no te that these monthly strong bonds occurred without much overlap between the females (Tables 4-8, 4-9). Grooming There was no signific ant monthly variati on in the rate (min/h) that females groomed troop-mates ( H = 17.775, df = 11, P = 0.087). There was, however, monthly variation in the rate at whic h females received grooming ( H = 41.689, df = 11,
128 P < 0.0001). Sample sizes were too small to te st statistically for monthly variation in each females dyadic grooming relationships. Patterns emerged, however, to suggest that the females primary adul t grooming partners did not remain consistent over time (Table 4-10), and were not effective indicators of monthly shifts in primary proximity partners; female primary pr oximity and grooming partners (o r lack thereof) matched only 35% of the time (Table 4-11). Grooming strong bonds were exhibited within 29 dyads: 14 and 15 pairs; only 13 dyads failed to exhibit a singl e strong (monthly) gr ooming bond during the length of the study, and nine of those dyads were never observed to groom. The five strongest bonds of the monthly grooming strong bonds occurred between (groomer and recipient), in descending order: (1) Gina and Banana (February), (2) Banana and Gina (February), (3) Gina and Boris (September), (4) Banana and Boris (January), (5) Jane and Boris (May), representing the relations hips between the two highest-ranking females and between the three highest-ranki ng females and the alpha male. The most consistent of the strong grooming bonds was a tie between Gina/Boris and Banana/Boris (each had 7 of 10 months characterized by st rong grooming bonds). The most reciprocal of the strong bonds, howev er, was the grooming relationship between Little Horns and Darwin; the dyad exhibited f our months of equit able, strong grooming bonds (Tables 4-8, 4-9). Of the strong bonds among females, Gi na and Banana expressed the most enduring and equitable bond over time; this relationship, however, did not extend beyond June. The majority of the strong bonds among females occurred in February. Of the strong bonds between a male and a female females (except for Little Horns)
129 tended to focus their grooming attenti on on the alpha male. Boris had a strong grooming bond with an average of three females during eac h month of the study, and these were highly inequitable grooming relations hips. In fact, Boris was responsible for only 15% of the monthly st rong bonds he held with females. Although Gina and Banana maintained the strongest grooming bonds with Boris, he did not maintain a single strong grooming bond with either of them. Andycap was never observed to invest in a grooming relationship with a female, but Darwin demonstrated at least one strong grooming bond with most of t he females. The most endur ing and equitable of strong bonds occurred between Little Horns and Darwin. Tinkerbell was the only female to not express a single strong grooming bond with the alpha male, nor did she invest in a grooming relationship with eit her of the subordinate males (Tables 4-8, 4-9). Seasonal Variation in Affiliation Proximity There was no signific ant seasonal variation in overall TFPS ( H = 2.668, P = 0.446, df = 3) or MPS ( H = 0.422, P = 0.936, df = 3). Female proxim ity to other females, however, was highest during the fruiting season (mean SE = 28.55 9.81) while female proximity to males was highest durin g the transition season (mean SE = 28.31 11.20). In comparing seasonal TFPS to MPS, females stayed closer to males than to females in each season except the frui ting season, although the only statistical difference was during the dry season, when fema les were significantly closer to males than to females ( Z = -3.702, P < 0.0001) (Figure 4-7). Grooming In all seasons, females groomed signific antly more than they received ( Zfruiting = 5.662, P < 0.0001; Zwet = -4.008, P < 0.0001; Zdry = -5.838, P < 0.0001;
130 Ztransition = -3.416, P = 0.001). The average rate that females groomed troop-mates was greatest (mean SE = 1.13 0.24 min/h) during the fruiting season (February to April), the time of year that corres ponded not only to fleshy fruit availability but also to the presence of young infants. Females gr oomed troop-mates the least (mean SE = 0.51 0.12 min/h) during the dry season. There was no significant seasonal variation, however, in the daily rate that females groomed conspecifics ( H = 4.718, df = 3, P = 0.194). Similarly, the average rate that females received grooming from troop-mates was greatest (mean SE = 0.54 0.18 min/h) during the fruiting season and smallest (mean SE = 0.12 0.04 min/h) during the dry season. There was significant seasonal variation in the daily rate that females received grooming ( H = 20.364, df = 3, P < 0.0001), and this variation was wholly attributed to decreased grooming during the dry season. Females rece ived a higher rate of grooming during the fruiting ( Z = -4.167, P < 0.0001) and wet ( Z = -3.727, P < 0.0001) seasons than during the dry season. Females devoted more grooming to females than to males only during the fruiting season, and this finding was nearly significant ( Z = -1.964, P = 0.05). In the remaining seasons, there was no statistical difference in the rate at which females groomed males versus females (Figure 4-8). There was no significant difference in any season in the rates at which female-female grooming occurred up versus down the hierarchy ( P = 0.109 to 0.786). Groomi ng up the hierarchy occurred more than the reverse direction only in the dry and transition s easons, with the greatest (up versus down) differential occurring in the tr ansition season (Figure 4-9).
131 The Effect of Infant s on Female Affiliation Proximity During the months in which infants ( n = 4) were 3 months old (January to April), there was no signific ant difference in female proximity to males (MPS) as compared to female proximity to females (TFPS) ( Z = -0.365, P = 0.715). When differentiating between females with infants (FwiPS) and fe males without infants (FPS), however, females were significantly more likely to be in close proximity to a female with an infant than to a female without an infant ( Z = -1.979, P = 0.048). Further, while MPS was significantly higher than FPS ( Z = -3.400, P = 0.001), there was no significant difference between MPS and FwiPS ( Z = -1.714, P = 0.086). This indicate s that females were more likely to be in close proximity to a male than to a female without an infant, but the same did not hold true when an infant was present. When only considering the proximity score s of the females wit h infants (Gina, Banana, Carol, and Kate), t heir total proximity score (T PS), including female, male, and immature neighbors, was significantly higher when they were carrying their young (0 to 3 month old) infants than when they were not with their infants ( Z = 2.714, P = 0.006). More specifically, when carrying their babies mothers had a significantly higher FwiPS ( Z = -2.135, P = 0.039) and MPS ( Z = -2.714, P = 0.007) than when they were not carrying their babies. Infant presence, however, did not affect their FPS ( Z = -0.420, P = 0.713) (Figure 4-10). Thus, the physica l presence of a young infant increased a females proximity to other females carrying infants and to males, but did not affect her proximity to females without infants. Further, while the mothers were carrying their young infants, their highest av erage proximity score was with other females with infants (mean SE = 32.2 11.10) as compared to males (mean SE = 24.4 7.19) or
132 females without infants (mean SE = 14.30 6.56). On the other hand, when the mothers were not carrying their young infant s, their highest average proximity score was with males (mean SE = 10.43 5.53) as co mpared to females without infants (mean SE = 5.58 4.37) and females with in fants (mean SE = 5. 00 3.29). Infants also appeared to affect the strong bonds that females formed with conspecifics. For example, other than Gina and Banana, the only other female dyad to form a strong bond that endured for more than one consecutive month was that of Banana and Kate. Their infants were t he youngest and closest in age (two-week age difference), and the strong proximity bonds between Banana and Kate occurred during the months of February and March when their infants were 1 to 2 mo nths old. Infant presence may have also been a factor infl uencing the strong prox imity bond between Gina and Banana that lasted from January to Ap ril. Interestingly, Gina, Banana, and Kate each demonstrated a strong proximity bond with Boris when their infants were approximately 0 to 1.5 months old. The relationship between Kate and Boris appeared particularly infant-based, both in their two-month strong proximity bond (the longest (consecutive) strong bond that any of the females with y oung infants had with the alpha male), and in the observation that Boris partici pated in the caretaking of Kates infant. In fact, Boris was the infants first obs erved alloparent, and he was regularly seen carrying the baby. Boris was not observed to carry any other infants. Carol was the only female with an infant that did not dev elop a strong bond with Boris; this may be due to the fact that Carols in fant was the eldest of the gr oups infants and was nearly two months old at the onset on the study. The strong bonds between Boris and the other mothers occurred when their infants were typically less than two months old, and
133 therefore it is possible that a similar relationship between Carol and Boris occurred prior to the study (Tables 4-8, 4-9). Grooming Females with infants ( FWI) [i.e., female s carrying and/or nursing an infant, usually but not necessarily, the mother] groomed cons pecifics for a total of 324.8 min. The largest percentage (35.7%) of this groomi ng was devoted to other FWI, while females without infants received the leas t (5.3%). Adult males receiv ed 26.9% of grooming from FWI, all of which was devoted to the alpha ma le; females carrying infants were never observed to groom a subordinate adult male. Females with infants received 274.2 min of grooming from conspecifics, the majori ty of which was given by other FWI and females (42.4% and 29.9%, respectively). A ll females except for Tinkerbell (the most subordinate female) were observed to groom FWI. Males (the alpha male and one of the subordinate males) least groo med FWI (3%) (Figure 4-11). During the months in which infants were approximately 0 to 3 months old (January to April), the grooming interactions of t he mothers (Gina, Banana, Carol and Kate) were affected by infant presence. When they ca rried their infants, the frequency and duration of grooming given versus received did not deviate from expected patterns ( X 2=0.06, df = 1, P > 0.10; X 2= 2.6, df = 1, P > 0.10, respectively); they groomed conspecifics during 76 bouts (177.8 min) and received from tr oop-mates 79 groomi ng bouts (209.6 min). When they were not carrying their infants, however, the females groomed troop-mates more than expected (26 bouts, 81.3 min) and received less grooming (7 bouts, 18.5 min) than expected ( X2 = 11.0, df = 1, P < 0.001; X 2=39.5, df = 1, P < 0.001, respectively). Although I did not collect data to compare the actual amount of time that females cared for their infant s with the amount of time that females spent away from
134 their infants, the female pr oximity scans provide such an estimate. From January to April, the mothers were recorded with their infants in 123 total scans and without their infants in 101 scans. Chi-square analysis rev eals that the frequency distributions do not significantly differ ( X2 = 2.16, df = 1, P > 0.10), allowing for a direct comparison within the categorical variable of infant presence. The fema les groomed more than expected when carrying their infants and less than expected when they were without their infants (number of bouts: X 2 = 26.75, df = 1, P < 0.001; duration (min): X2 = 56.48, df = 1, P < 0.001). Similarly, the females received grooming more than expected when carrying their infants and less than expected when they were without their infants (number of bouts: X2 = 57.36, df = 1, P < 0.001; duration (min): X2 = 129.05, df = 1, P < 0.001). In fact, the females gave nearly three times more grooming and received approximately 10 times more grooming when they were carrying their infants (Figure 4-12). Infants also appeared to affect the strong (grooming) bonds t hat females formed with conspecifics. For example, all of the grooming comprisi ng Tinkerbells strong bonds with Banana and Kate in April occurred while she was carrying Bananas infant. Otherwise, Tinkerbell was nev er observed to groom or be groomed by a female. Additionally, all of the females with inf ants maintained strong gr ooming bonds with the alpha male for at least the fi rst month (and generally for the first three months) of their babies lives, while the females without infants did not typically form strong bonds with Boris during this same time. The only male to invest in a relationship with a FWI was Darwin grooming Kate in t he first month of her infant s life (Tables 4-8, 4-9). Discussion Are the female brown capuchins in RV truly female-bonded? Ac cording to the common indices of grooming and proximit y measured on an annual time-scale, the
135 short answer is no; the females are not fema le-bonded in the classic sense and instead are better characterized as cross-bonded with males. Despite a 2:1 female to male ratio, females groomed males significant ly longer than they groomed females, and female proximity scores with males were si gnificantly higher than female proximity scores with other females. Further, each i ndividual females highest average proximity score (aside from Bananas) and longest tota l grooming duration was with a male, and not with a female. For the most part, pr oximity and grooming measures resulted in similar strong bonds which emp hasized female relationships with males. The only exception was the relations hip between Gina and Banana, the two highest ranking females, although each of them was more str ongly bonded to a male than to each other. Also of interest, proximity to males increas ed with female rank, while there was no clear pattern between female rank and proximity to females Based on this evidence, male C. apella in RV can be viewed as a valuable resource to the females, and the social importance of males may be related to the relatively weak relationships between females (Izar 2004). Especially telling is the finding that the strongest bonds between females were through proximity, while the strongest female-male bonds were through groomi ng. This indicates that the females were not simply grooming troop-mates that were closest or most available to them. Instead, they actively searched out male gr ooming partners. Females are expected to be affiliative to those individuals that provid e the greatest benefits; in my study, patterns of affiliation suggest that males, and in particular the alpha male, provided the most potential benefits to females (O 'Brien 1991). OBrien (1991) suggested that preference for the alpha male is a reflection of breeding preference, while Janson (1985) suggested
136 that female C. apella may preferentially groom the dominant male in exchange for his tolerance. The benefits t hat female brown capuchins In RV may accrue through intensified grooming interactions with the alpha male are considered below. In sum, on an annual-time scale, it was cl early evident that, despite being female philopatric, the females in this group we re not female-bonded. But was there within-year variation in female relations hips? Understanding short-term patterns of affiliation may provide insight into the functional significance of non-random dyadic social behavior. Outcome of Hypotheses Social bonds between adult female brow n capuchins are variable over time This hypot hesis is supported. Both m easures of grooming and proximity revealed substantial monthly variation in each fe males primary partner, and given a larger sample size, affiliative partners likely sh ift on an even shorter-term basis. Females demonstrated 60% inconsistency in their mont hly primary partners (both in terms of proximity and grooming), and strong bonds t hat were strong through both grooming and proximity typically lasted for only one month. Such a strong bond, however, endured for three consecutive months between Gina and Banana. These two females were the highest-ranking females, and it is speculated that they are a mother-daughter dyad. Silk et al. (2006a) determined t hat among female baboons, bond strength was a decreasing function of relatedness, with females forming the strongest bonds with mothers and sisters. Once genetic analysis of the females in my study is complete, a story may emerge that is both explanative and puzzling. Maternal relatedness may explain the relationship between Gina and Banana, but the obvious lack of
137 female-female bonds within most dyads is all the more intriguing given the high degree of relatedness that is ex pected among the females. A females primary proxim ity partner did not necessa rily predict her primary grooming partner for that same month, and vice versa. In fact, only 35% of cases matched, and monthly dyadic grooming rate s and proximity scores were only modestly correlated. This brings to light two me thodological questions: (1) Are grooming and proximity equivalent measur es of a bond? (2) Should a bond be assigned to a dyad only when both measures agree? Grooming and prox imity may not necessarily be equitable proxies of a bond, but are likely compatible and together more fully expose the facets of a dyadic relationship. I acknowledge a desig n flaw in my data collection that did not allow me to assign which member of a dyad was responsible for maintaining proximity to the other. Therefore, for my study, I place more explanative power in grooming behavior as it indicates an individuals willing ness to invest in an affiliative relationship with another. The comparat ive and corroborative pattern s that emerged through the proximity data, however, are still worthy of inclusion and cautious consideration. Dyadic female relationships vary in resp onse to ecological pressures Support for this hypothesis is tentative. Seasonal differences in social behavior were not consistently supported by statis tics, possibly because the generalized diet of capuchins provides them relative independence from ecological constraints (Moura 2007). Suggestive patterns emerged, however to indicate the potential benefits provided by dyadic association. Female proximity to males and investment in the grooming of males is most evident during the transition season, a time of year characterized by an unpredictable food supply and a subsequent need to increase travel rates, increasing exposure to predator s (Chapters 2 and 3). In support of Janson
138 (1985), this could be viewed as a social strategy whereby the females increase grooming of the males, and par ticularly the alpha male, in exchange for tolerance at food sources and during travel. Evidence of such a possible exchange was actually provided by female-female grooming patterns. While not statistically significant, females groomed up the hierarch y more than the reverse di rection only during the dry and transition seasons. This could be interpreted as a female strategy to groom higher ranking troop-mates during periods of low and/or unpredictable food availability in attempt to procure tolerance at food source s. Direct evidence for such an exchange, however, was not provided by my study. However, females did not appear to use grooming as a commodity to be exchanged for tolerance when food was r eadily and, possibly more importantly, predictably, available. For example, fema les received the most aggression from males, especially the alpha male, during the wet season when feeding on M. maripa a high-quality resource that fruits in monopo lizable clusters (Chapter 2). Rates of female-male grooming during this time, however, were virtually the lowest among the four seasons. Therefore, although female s do not appear to use grooming as a means to reduce aggression from males, females ma y utilize the grooming of males as a form of insurance during periods of instability and/or unpredictability, securing better access to resources (tolerance during feeding) and decreasing risk (improving predator protection via closer proximity to males) Hierarchical grooming among females may serve as a commodity to be tr aded for tolerance at food res ources; investigation into such a fine-scaled dyadic transaction deserves further study.
139 Females stayed closer to females and devot ed more grooming to females than to males only during the fruiting season. Upon init ial consideration, this increase in female relationships may appear to be in response to the peak production of fleshy fruits, possibly as a means to increase intrasexual to lerance while feeding. If this were the case, however, I would expect grooming during th is time to be directed up the hierarchy (Seyfarth 1977), when, in fact the rate at which females groomed up the hierarchy was slightly lower than the rate at which fema les groomed down the hier archy. While the association between female relationships and fr uit availability should not be ruled out in entirety, the intensificati on of female-female behavior dur ing the fruiting season was confounded by the presence of in fants, as four of the sev en females had infants that were less than three months old during this time. Relationships among females strengthen w ith the presence of an infant This hypothesis is supported, indicating that young infants were a source of interest to troop-mates, and that, with the presence of an inf ant, a mother became more socially immersed. In terms of monthly strong bonds between females, especially those that were strong through both grooming and proximit y, 78% occurred when at least one of the females had a young infant; t he females that did not have infants did not show a tendency to form strong bonds duri ng this time. Therefore, despite the dynamic and seasonally variable environment, infant presence seemed to most affect female-female relationships. For example, other than Gina and Banana, the only other female dyad to form a strong bond that endured for more than one consecutive month was that of Banana and Kate Their infants were the youngest and closest in age (two-week age difference), and the strong bonds between Banana and Kate occurred during the months of February and March when their infant s were 1 to 2 months old.
140 Infant presence may have also been a factor influencing the strong bonds between Gina and Banana that lasted from January to Apr il, as both females had young infants during this time, and Gina was not obs erved to groom Banana after May. Thus it is evident that young infants were the primary cata lyst for the intensification of female-female relationships. The question then becomes, Why? Anecdotal evidence suggests that the most subordinate fema le, Tinkerbell, ex changed alloparenting services for affiliative behavior from females: All of the grooming comprising Tinkerbells strong bonds with Banana and Kate in April occurred while she was carrying Bananas infant; otherwise, Tinkerbell was never obser ved to groom or be groomed by a female. However, this occurred only over a two-day period, and the use of alloparenting as a commodity to be traded (for re ciprocal alloparenting or for grooming), did not appear to be widespread among the females in this group. For example, Banana frequently babysat Ginas infant, but the reverse wa s rarely, if ever, observed. In fact, alloparenting among the females (mother s and non-mothers) was not commonly observed; instead it was the juveniles (male and female) that mo st frequently carried infants. I find it more likely that t he intensification of proximit y and grooming relationships among females from February to April was an artifact of mothers becoming more central in the group, as the females we re nearly twice as likely, on average, to be located in the center of the group wh en they had infants that were less than three months old. Whether the females became more central as a means to increase protection from predators and access to higher quality food resources or to socialize their infants is unknown. Often, however, it appeared that new mother s (especially the more
141 subordinate ones) escaped, by choice, to the periphery to forage whenever possible as though overwhelmed by their sudden popularity As grooming behavior promotes the release of -endorphins (Keverne et al. 1989) and thus reduces tension, it is my belief that a new mother intensifies grooming behavior, in part, as a means to cope with the social stress she incurs while socializing and protecting her infant (to be empirically examined in the follo wing chapter). Female-Male Relationships The results of this chapter indicate that, overall, females were more motivated to bond with males, and in particular the alpha male than with other females. This general pattern of female-male affiliation resemble s that which is demonstrated among some hamadryas baboon populations ( Papio hamadryas ), w hereby each female has a stronger bond with the dominant male than with the other adult females (Byrne et al. 1990; Barton et al. 1992; Swedell 2002). Su ch populations with strong inter-sexual affiliative relationships are characteriz ed as cross-bonded (Byrne et al. 1990), a term that has not previously been used to descr ibe any capuchin group or population. Among C. apella and C. olivaceus it is well-established that, of males, females both sexually and socially prefer the dominant male (Janson 1984; Robinson 1988; O'Brien 1991; Di Bitetti 1997; Izar 2004). Close association with the dominant male, however, is not without its cost s. In fact, nearly 70% of the agonism females received in my study came from the alpha male, and t he two highest-ranking (and most central) females received approximately 2 to 3 time s more agonism from the alpha male than did the other females. The benefits afforded by interaction with the dominant male, then, must be great enough to outweigh the costs. Estrus -cycling females and females
142 with infants are expected to receive different benefits from their association with the dominant male, and I discuss these differ entiated relationships below. Associations between females w ith infants and adult males All of the females with infants mainta ined strong grooming bonds with the alpha male, Boris, for at least the first month (and generally for the first th ree months) of their babies lives. The relationship between Kate and Boris appeared particularly infant-based, both in their two-month str ong (grooming and proximity) bond (the longest consecutive strong bond that any of the females with y oung infants had with the alpha male), and in the observation that Boris partici pated in the caretaking of Kates infant. In fact, Boris was the infants first obs erved alloparent, and he was regularly seen carrying the baby. Boris was not observed to carry any other infants. Carol was the only female with an infant that did not dev elop a strong bond with Boris; this may be due to the fact that Carols in fant was the eldest of the gr oups infants and was nearly two months old at the onset on the study. The strong bonds between Boris and the other mothers occurred when their infants were typically less than two months old, and therefore it is possible that a similar relationship between Carol and Boris occurred prior to the study. The mothers were clearly responsible for maintaining the close association with the alpha male (as he rarely reciprocated grooming) and it is of intere st to note that, of the adult males, females with babies only gr oomed Boris. In this multi-male, multi-female species, infant paternity is uncertain as estrus females typically mate with more than one male. The alpha male, howev er, generally has the most access to females in estrus at the ti me they would be ovulating (Janson 1984) and so likely sires the majority of the groups in fants. Preferential mating a nd post-natal association with
143 the dominant male may be a strat egy to protect infants from infanticide. However, the females did not appear especially wary of the subordinate males and, in fact, Darwin, the youngest adult subordinate male, demonst rated an unreciprocated strong grooming bond with Kate when her infant was less than one-month old (i.e., Darwins advances were tolerated by the mo ther of a newborn). The benefits accrued from a mothers association with the alpha male are not mutually exclusive. As the dominant male occupies a central position in the group and controls access to resources (Janson 1985; Janson 1990a,b), it is likely that a mothers affiliative investment in such a relations hip garners a suite of benefits in addition to protection from potentially infanticidal males, such as: increased tolerance at high-quality food resources (especially import ant for meeting the energetic demands of lactation) and in the center of the group (i.e., increased predat or protection), the security of his close proximity (i.e., protection for her and her infant from agonism and/or bothersome conspecifics), and the introduction and socialization of her infant with the dominant male, the most important individ ual in the group (i.e., preparing for her offsprings future). Associations between reproductivelycy cling females and adult males Of the three females that did not have infants during the study period, Jane and Tinkerbell demonstrated the stro ngest and most frequent affiliative relationships with the alpha male. Little Horns, however, was argu ably more closely bonded to Darwin than she was to Boris, although she demonstrat ed four months of strong bonds to both males. The relationships of cycling fema les with males may be a reflection of their breeding preference; however, as affiliative behavior observed during a females estrus cycle was not included in analysis, the resu lting patterns do not represent a direct
144 connection between female-male affilia tion and mating. Further, among C. apella mate preference occurs through female choice, and so it is unlikely that a female would need to use grooming, for example, to entice a male to mate with her. I find it more likely that these females preferentially associate with the dominant male as a means to acquire the benefits of his tolerance, namely access to high quality resources and improved predator protection. Males, however, focused their affiliative behavior on the female s that did not have infants, a finding that I find particularly in teresting given the level of motivation new mothers demonstrated in associating with the dominant male. For example, from January to June when the inf ants were < 6 months old, the alpha male groomed non-mothers ( n = 3) more than twice as long as he groomed mothers ( n = 4), although the females with infants groomed him mo re than four times longer than did the non-mothers. The relationships between t he alpha male and females with infants were decidedly one-sided, with males demonstrating a clear preference for the females that were most likely to conceive investing in their future reproductive success at the possible expense of their current reproducti ve output. Again, this preference was quantitatively obvious even after excluding observations made during the females estrus periods. Because females exercise mate preference, and the females had their choice of three possible mates in the gr oup, males appear to utilize grooming behavior as a means to secure access. The fact that the dominant ma le was just as, if not more, active than the subordinate males in mainta ining affiliative contact with the cycling females suggests that he cannot rely solely on his status to acquire matings.
145 Conclusion The concept of female bonds is at the core of theories of primate social evolution, and compelling evidence suggests that females sharing strong bonds obtain significant survival and reproductive advantages. T he obvious benefits of female-female bonds, therefore, make the lack of such relationships all the more in triguing. The results of my study suggest that the females in a femalephilopatric species may not necessarily view other femal es as their most beneficial social partners. Female brown capuchins in Raleighvall en were not female-bonded in the classic sense. Instead, the females were ar guably cross bonded especially with the alpha male. This was clearly evident when analyzing the data on a yearly timescale. Finer-scaled temporal analysis, however, revealed that the females were sometimes female-bonded. Long-term social bonds, therefor e, may not be necessary, or even beneficial, in a dynamic ecological landscape. Female relationships did not appear to be affected by seasonal food availability. Although a hierarchy among the female s was weakly evident, there was a noticeable lack of interaction among most females. Because food patches in Raleighvallen are either very small or very large, and the monkeys rely heavily on insects, females generally did not affect each others access to food, and this may result in a lack of mo tivation for females to bond. Instead, affiliation between females was centered on infants. Lactating females are more sensitive to fluctuations in food availability, and so infants may intensify ecological pressures for their mothers and encourage stronger female relationships. Alternatively, infant pr esence may encourage a female to bond with conspecifics as a means to socialize or protect the infant, and this may explain the strong bonds that the mo thers formed with the alpha male. Mothers may also increase affiliation as a means to cope with the physiological and social burden of caring for, protecting, and socializing their infants. In sum, I believe that the distinction bet ween friends and business partners is better viewed as the social endpoints on an ecological continuum. In a dynamic environment, dynamic social relationships ar e most adaptive, with the true value of a social bond possibly within its potential for mobilization (Henzi et al. 2009). As such, relationships between female capuchins in RV can be described, at best, as transient
146 and seemed to mainly provide dyadic benef it when an infant was born. Instead, relationships with adult males, particularly the alpha male, appeared to provide the females with more long-term benefits likely a ssociated with mediating the costs of high predation risk and an unpredictable food supply.
147 Table 4-1. Summary of the dominance rank approximate age, and r eproductive history of the adult females in Troop A, as well as general life history notes for the adult males. Females Rank Est. age (yrs) Birthing history (most recent birth as of Dec. 2006) Notes of interest Gina 1 > 10 Multiparous (Dec. 24, 2005) Banana 2 6 Primiparous (Jan. 21, 2006) Jane 3 8 Multiparous, pregnant (Feb. 22, 2005) Disappeared in October Suspected obstetric-related death (near full term) Little Horns 4 > 15 Multiparous (Dec. 6, 2004) Carol 5 > 10 Multiparous (Nov. 22, 2005) Kate 6 > 10 Multiparous (Feb. 5, 2006) Tinkerbell 7 6 Nulliparous, pregnant Gave birth to 1st infant in December Males Boris alpha > 10 Alpha male of Troop A since 2003 Andycap > 15 Immigrated into Troop A in January 2006 Darwin 8-10 Immigrated into Troop A in April 2005; youngest adult male in troop
148 Table 4-2. Female dominance matrix depicti ng a (weakly) linear hierarchy based on the outcomes of dyadic agonism, transitivity, and anecdotal evidence. Gina Banana Jane Little HornsCarol Kate Tinkerbell Gina 1 1 1 ^1 1 1 Banana *1 1 1 1 ^1 Jane 1 1 ^1 1 Little Horns 1 ^1 *1 Carol 1 *1 Kate *1 Tinkerbell 1 = row individual is dominant to co lumn individual. ^1 = relationship based on transitivity. *1 = relations hip based on anecdotal evidence. Table 4-3. Average monthly proximity scores ( SE) of Troop A females. The higher the proximity score, the more ti me the female spent in close proximity (within 4 m) to those conspecifics. Proximity scores were calculated using weighting factors such that proximity of 1 m provided a larger sco re than did proximity of 4 m (adapted from Perry 1996). Overall, all females except the alpha female, Gina, spent more time in closer proxim ity to adult males than to adult females despite a 2:1 ratio of potential female and male neighbors, respectively. Two females, Little Horns and Tinkerbell, produced proximity scores with males that were significantly hi gher than their proximity sco res with other females. Female proximity to males increa sed linearly with female rank. Proximity to: Female Rank Adult females ( n = 6) Adult males ( n = 3) P -value Gina 1 52.7 15.47 41.1 11.54 0.721 Banana 2 38.4 13.10 39.3 11.83 0.878 Jane 3 5.6 2.37 21.5 10.72 0.091 Little Horns 4 1.5 0.79 14.3 3.69 0.017* Carol 5 10.7 4.91 12.4 3.64 0.767 Kate 6 11.6 5.83 12.2 5.90 0.859 Tinkerbell 7 0.7 0. 68 9.5 4.96 0.008*
149 Table 4-4. Frequency and duration of female grooming dyads (by age/ sex class), from January to December 2006. Females mo st groomed juveniles (36% duration) and least groomed females (19% duration). Average bout length was longest when females groomed males. Females received the most grooming from females (49% duration), and not incl uding infants, received the least grooming from males (17% duration). Groomer Recipient Total # bouts Total duration (min) X bout duration SE (min/bout) Females Males 140 429.1 3.1 0.27 Females Juveniles 301 700.2 2.3 0.15 Females Infants 201 437.7 2.2 0.16 Females Females 175 370.8 2.1 0.17 Males Females 62 130.0 2.1 0.32 Juvenile Females 126 246.0 2.0 0.16 Infants Females 8 8.3 1.0 0.35
150 Table 4-5. Total proximity scores for adult dyads, January to December 2006. Shaded scores represent strong bonds* between individuals. The only st rong proximity bond observed between fe males (Gina and Banana, the two highest ranking females) was also an equitable one. The remaining strong bonds occurred between the females and the alpha male, Boris. Gina Banana Jane LH Carol Kate Tink Boris Andycap Darwin Gina X 233.69 25.55 18.62 23.41 31.77 1.55 272.49 14.10 53.14 Banana 233.27 X 28.16 0.21 10.11 54.40 8.536 151.67 46.41 51.86 Jane 43.93 6.89 X 2.62 4.37 4.27 0.00 114.05 4.30 82.56 Little Horns 6.80 0.70 1.87 X 5.78 1.84 1.70 41.65 23.01 50.60 Carol 21.35 10.11 4.37 5.78 X 46.10 1.36 23.85 49.84 48.46 Kate 31.77 54.40 4.27 1.84 46.10 X 0.73 83.93 3.79 19.97 Tink 1.55 11.15 0.00 1.70 0.50 0.73 X 110.03 14.30 2.56 Notes: Row individual = female for which proximity scores were calculated. Column indivi dual = neighbor; Strong bonds are those which comprise the top 10% of average proximity scores among adult dyads.
151 Table 4-6. Total grooming duration (min) for adult dyads, J anuary to December 2006. Shaded scores represent strong bonds* between individuals. The only strong grooming bond observed between females (Gina and Banana, the two highest ranking females) was also an equitable one (i.e., a strong bond for each partner in the dyad). The only other equitable strong grooming bond occurred within a dyad, Little Horns (L H) and Darwin. All other strong grooming bonds involved a female (inequitably) grooming the alpha male, Boris. Gina Banana Jane LH Carol Kate Tink Boris Andycap Darwin Gina X 83.8 1.0 2.0 3.5 1.0 0.0 86.5 0.0 0.0 Banana 49.8 X 0.0 3.5 0.0 18.8 4.7 82.5 0.0 3.0 Jane 12.0 12.2 X 17.8 2.3 0.0 0.0 36.3 0.0 0.0 Little Horns 12.0 4.0 3.5 X 8.5 6.0 0.0 7.1 0.5 40.6 Carol 4.5 9.0 3.2 9.2 X 2.0 1.0 24.2 10.0 20.0 Kate 0.0 17.5 0.0 2.0 0.2 X 3.0 51.2 5.3 0.0 Tink 0.0 2.5 0.0 0.0 1.0 0.0 X 0.0 0.0 0.3 Boris 3.7 6.7 18.2 16.2 5.0 6.1 0.0 X X X Andycap 0.0 0.0 0.0 1.3 16.0 0.0 0.0 X X X Darwin 5.3 0.0 2.0 30.2 1.5 5.0 7.0 X X X Notes: Row individual = groomer. Column individual = recipient; Strong bonds were those which comprise the top 10% of grooming duration among adult dyads
152 Table 4-7. Monthly variation in females primary adult proximity partners, based on the highest dyadic proximity score/month (shown). Female proximity partner s did not remain consistent over time. Jan Feb Mar Apr May Jun Aug Sept Oct Nov Gina Boris 47.34 Banana 44.85 Banana 16.84 Banana 32.79 Boris 24.85 Banana 27.27 Boris 75.75 Boris 37.89 Darwin 17.07 Boris 57.31 Banana Gina 45.71 Gina 44.85 Boris 28.42 Gina 32.79 Darwin 17.18 Gina 27.29 Gina 25.69 Gina 17.85 Andycap 15.73 Darwin 7.77 Jane ----Gina 19.35 --Boris 5.06 --Boris 4.24 Boris 44.62 --n .a. Little Horns --Darwin 17.88 Boris 14.37 Andycap 3.68 Darwin 7.56 Darwin 16.77 Andycap 4.50 Boris 4.50 Boris 16.77 Carol 3.11 Carol Kate 1.20 Gina 8.03 Kate 22.41 --Gina 13.32 ----Darwin 31.73 Boris 18.20 Little Horns 3.11 Kate Carol 1.20 Banana 19.91 Boris 35.51 --Boris 22.75 Boris 4.00 Gina 16.37 --Boris 3.33 --Tinkerbell ------------Andycap 6.80 ----Boris 4.86 Notes: July and December were not incl uded due to insufficient data; only dyads with > 1 instance of proximity (within 4 m) were considered.
153 Table 4-8. Monthly variation in proximity (p) and grooming (g) strong bonds, comprising the top 10% of monthly proximity scores and grooming durations (such that the focal female is the groomer, dyadic partner the recipient) among adult female dyads. Only 20% of strong proximity bonds matched the strong grooming bonds (indicated by pg). The majority (80%) of bonds that were strong though both grooming and proximity, occurred in the presence of young infants ( 3 months old), as indicated by shaded cells. Of the 21 dyads, only three demonstrated more than one month of reciprocal strong bonds. Notes of interest: Gina and Banana had the most enduring and equitable of strong bonds, but this relationship weakened after April and altogether disappeared after September; Kate and Banana had the next most equitable relationship, but only when their infants were 1 to 2 months old; Jane and Little Horns had a fairly consistent, albeit inequitable, relationship based in grooming. Focal female Dyadic partner Jan Feb Mar Apr MayJun Aug Sep Oct Nov Gina Banana Banana Gina p p pg pg pg pg pg p g p pg p p p p Gina Jane Jane Gina pg p p Gina Little Horns Little Horns Gina g g Gina Carol Carol Gina g g Gina Kate Kate Gina p p Gina Tinkerbell Tinkerbell Gina Banana Jane Jane Banana g p Banana Little Horns Little Horns Banana g Banana Carol Carol Banana g Banana Kate Kate Banana pg pg pg p g g Banana Tinkerbell Tinkerbell Banana g
154 Table 4-8. Continued. Focal female Dyadic partner Jan Feb Mar Apr MayJun Aug Sep Oct Nov Jane Little Horns Little Horns Jane g g g g g Jane Carol Carol Jane g g Jane Kate Kate Jane Jane Tinkerbell Tinkerbell Jane Little Horns Carol Carol Little Horns g g g g Little Horns Kate Kate Little Horns g Little Horns Tinkerbell Tinkerbell Little Horns Carol Kate Kate Carol p p Carol Tinkerbell Tinkerbell Carol Kate Tinkerbell Tinkerbell Kate g
155 Table 4-9. Monthly variation in proximity (p) and grooming (g) strong bonds, comprising the top 10% of monthly proximity scores and grooming durations among adult dyads. Bonds strong through proximity were only measured from the focal female perspective. Grooming strong bonds are indicated for both dyadic partners, such that the focal individual was the groomer and the dyadic partner the recipient. Only 18% of a females strong proximity bonds matched her strong grooming bonds (indicated by pg). Of bonds that were strong though both grooming and proximity, 45% occurred between a female with a young infant and the alpha male, Boris, as indicated by shaded cells; 64% occurred between Boris and the two most dominant females (Gina and Banana). Of the 21 dyads, only two demonstrated more than one month of reciprocal strong grooming bonds. Notes of interest: Gina and Banana maintained the most enduring strong grooming bonds with Boris, although he did not invest in a grooming relationship with either of them. Andycap, the eldest subordinate male, was not observed to invest in grooming relationships with any females. Darwin, on the other hand, demonstrated at least one month of strong grooming bonds with most females. The most enduring and equitable of grooming relationships was that of Little Horns and Darwin. Focal individual Dyadic partner Jan Feb Mar Apr May Jun Aug Sep Oct Nov Gina Boris Boris Gina pg g g pg p pg g pg Banana Boris Boris Banana pg g pg p p pg g g g Jane Boris Boris Jane p g g g g g g p g Little Horns Boris Boris Little Horns p g g g p Carol Boris Boris Carol g g g g g p Kate Boris Boris Kate g pg pg g p g Tinkerbell Boris Boris Tinkerbell p p p Gina Andycap Andycap Gina Banana Andycap Andycap Banana p p
156 Table 4-9. Continued Focal individual Dyadic partner Jan Feb Mar Apr May Jun Aug Sep Oct Nov Jane Andycap Andycap Jane Little Horns Andycap Andycap Little Horns Carol Andycap Andycap Carol p p g Kate Andycap Andycap Kate g Tinkerbell Andycap Andycap Tinkerbell Gina Darwin Darwin Gina g p p Banana Darwin Darwin Banana p p g Jane Darwin Darwin Jane p p g Little Horns Darwin Darwin Little Horns pg g g g g g pg g g Carol Darwin Darwin Carol p p g Kate Darwin Darwin Kate g p Tinkerbell Darwin Darwin Tinkerbell g
157 Table 4-10. Monthly variation in females primary adult grooming partner, based on the highest dyadic grooming rate (min/h) per month (shown). Fe male grooming partners did not remain consistent over time. Jan Feb Mar Apr May Jun Aug Sept Oct Nov Gina Boris 0.085 Banana 0.387 Banana 0.120 Boris 0.121 Banana 0.042 Banana* 0.018 -Boris 0.081 Boris 0.038 Boris 0.051 Banana Boris 0.110 Gina 0.387 Gina* 0.120 Tink 0.034 Gina 0.042 Gina 0.018 Boris 0.049 Kate* 0.044 Boris 0.106 Boris 0.083 Jane -Banana 0.060 Gina 0.036 Lit. Horns 0.016 Boris 0.146 Boris 0.025 Boris 0.018 Lit. Horns 0.021 Boris 0.143 n.a. Little Horns Gina 0.040 Darwin 0.085 Banana 0.009 Darwin 0.107 Darwin* 0.073 Darwin 0.055 -Carol 0.032 -Banana 0.027 Carol Boris 0.020 Boris 0.055 Banana 0.027 Boris 0.045 Andycap0.102 Banana 0.009 -Lit. Horns 0.032 Andycap0.014 Darwin 0.095 Kate Boris 0.080 Banana 0.067 Boris 0.059 Boris 0.112 Boris 0.027 --Banana 0.044 Andycap0.021 Banana 0.011 Tink -Darwin* 0.040 -Banana 0.034 ---Carol 0.008 --Notes: July and December were not included due to insufficient data; only dyads with > 1 min of grooming were considered. Indicates the dyadic groom ing relationship was maintained primarily by the partner; otherwise, the dyadic relationship was relatively equitable or ma intained primarily by the focal female.
158 Table 4-11. Females primary adult partner per month (proximity/grooming). The s haded cells are the months in which a females primary grooming partner (or lack thereof) was the same as her pr imary proximity pa rtner. Grooming and proximity partners matc hed in 35% of cases. Jan Feb Mar Apr May Jun Aug Sept Oct Nov Gina Boris Boris Banana Banana Banana Banana Banana Boris Boris Banana Banana Banana Boris -Boris Boris Darwin Boris Boris Boris Banana Gina Boris Gina Gina Boris Gina Gina Tink Darwin Gina Gina Gina Gina Boris Gina Kate AC Boris Darwin Boris Jane ---Banana Gina Gina -LH Boris Boris -Boris Boris Boris Boris LH -Boris n.a. Little Horns -Gina Darwin Darwin Boris Banana AC Darwin Darwin Darwin Darwin Darwin AC -Boris Carol Boris -Carol Banana Carol Kate Boris Gina Boris Kate Banana -Boris Gina AC -Banana --Darwin LH Boris AC LH Darwin Kate Carol Boris Banana Banana Boris Boris -Boris Boris Boris Boris -Gina --Banana Boris AC -Banana Tink ---Darwin ---Banana ----AC --Carol --Boris -
159 Kin-based Figure 4-1. Deconstructing intra-sexual soci al bonds: Various usages of the term social bond and its synonyms, as found in the primate literature. Published usage of a term is included only if a definiti on or descriptive measure was provided, no matter how vague. 1Wrangham 1980; 2Mitchell et al. 1991; 3OBrien 1993; 4Di Bitetti 1997; 5Hemelrijk and Luteijn 1998; 6Manson et al. 1999; 7Payne et al. 2003; 8Eisenberg et al. 1972; 9Palombit et al. 1997; 10Silk 2002; 11Byrne et al. 1990; 12van Schaik 1989; 13Perry 1996; 14Henzi et al. 2000; 15de Waal 1986; 16van Hooff and van Schaik 1994; 17Sambrook et al. 1995; 18van Schaik and Aureli 2000; 19Nakamichi and Shizawa 2003; 20Jack 2003; 21Izar 2004; 22Silk 2007b Social Bond Pair Bond: relationship based on mutually reinforcing activities (e.g., daily grooming and other non-sexual behavior)8 Not (necessarily) kin-based Female-bonded: close affiliative link11; agonistic alliances12; long-term, differentiated, affiliative relationships based on grooming, proximity, dominance interactions, and coalitionary aid13; females form primary adult relationshi p s with females, based on g roomin g 14Obvious distress if individuals are separated, expend energy to reunite, show positive behavior upon reunion15 Cooperative relationships and mutual attraction based on proximity, tolerance, and affiliation 16 Intense grooming17,19,21 More time in close proximity and grooming, lower rates of agonism, more mutual agonistic support than the average dyad18 Affiliative relationship20 Close and stable relationship based on grooming22 Friendship: extraordinary grooming or proximity9; frequent affiliation, coalitionary support, mutual proximity, reciprocity, and tolerance10 Female-bonded=female philopatric. Differentiated grooming and agonistic support1; affiliation2; grooming3,4,7; reciprocal grooming 5; affiliation and cooperation6.
160 Figure 4-2. Total duration (min) fema les groomed and received grooming from adult male and female conspecifics. The am ount of time females spent grooming females was in positive association with fe male rank (descending from left to right). Similarly, aside from Jane (who disappeared durin g the study), the amount of time females invested in grooming males was (approximately) positively associated with female rank. The amount of time that females received grooming from adults, howeve r, does not appear associated with rank. 0 20 40 60 80 100 120 140 GinaBananaJaneLittle HornsCarolKateTink Groom Groom Receive grooming from Receive grooming from Total dyadic grooming duration (min)
161 Figure 4-3. The monthly rate (min/h) at which females groomed up the hierarchy as opposed to down the hierarch y. There was no significant difference in the rate at which subordinate fe males groomed dominant females (X = 0.030 0.008 min/h) and dominant females groomed subordinate females (X = 0.029 0.009 min/h) (P = 0.909). 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 Groom upGroom downX monthly rate of grooming (min/h)
162 Figure 4-4. Total duration (min) that adult females groomed troop-mates grouped by age/sex class (black bars), and the total duration that adult females received grooming from troop-mates grouped by age/sex classes (grey bars). Females invested the most time in grooming juveniles. Females groomed males more than females groomed female s; in general, however, the time investment of female-female grooming was more reciprocal than was femalemale grooming. 0 100 200 300 400 500 600 700 800 MalesFemalesJuvenilesInfantsGroomin g duration ( min )
163 Figure 4-5. Total duration and average bout length of and grooming. Females groomed males (grey bars) for a longer total duration than females groomed females (black bars), and grooming bout s were significantly longer ( P = 0.022) when females groomed males as opposed to female-female grooming bouts. 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 Total duration (h)X bout length (min) P = 0.022 Groomin g duration
164 Figure 4-6. Association between total prox imity scores and grooming durations (min), including all female-female and female-male dyads ( rs = 0.41, P = 0.001, n = 63). -10 0 10 20 30 40 50 60 70 80 90 -5050150250Grooming duration (min)Proximity score
165 Figure 4-7. Seasonal variation in female proximity to female s (black bars) and to males (grey bars). Female proximity to other females was highest during the fruiting season, while female proximity to males was highest during the transition season. Females stayed closer to males than to females in each season except the fruiting season, although the only statistical difference was during the dry season, when females were signi ficantly closer to males than to females ( P < 0.0001). Figure 4-8. Seasonal variat ion in the rate at which females groomed females (black bars) and females groomed males (grey bars). Females devoted more grooming to females than to males only during the fruiting season, and this finding approached significance ( P = 0.05). In the remaining seasons, females groomed males more than females although there was no statistical difference in the rate at which females groomed males as opposed to females. 0 0.1 0.2 0.3 0.4 0.5 0.6Fruiting Wet DryTransition 0 5 10 15 20 25 30 35 40 45Fruiting Wet Dry TransitionProximity score (X SE) Fruiting (Feb-Apr) Wet (May-Jul) Dry ( Au g -Oct ) Transition ( Nov-Jan ) ( P < 0.0001) X groom rate (min/h) SE Dry ( Au g -Oct ) Transition ( Nov-Jan ) Wet (May-Jul) Fruiting (Feb-Apr) ( P = 0.05)
166 Figure 4-9. The monthly rate (min/h) at which females in each season groomed up the hierarchy (black bars) as opposed to dow n the hierarchy (grey bars). There was no significant difference in any s eason in the rates at which femalefemale grooming occurred up ve rsus down the hierarchy ( P = 0.109 to 0.786). Grooming up the hierarchy occurred more than the reverse direction only in the dry and transition seasons, with the greatest (up versus down) differential occurring in the transition season (Figure 4-9). Figure 4-10. Proximity of mothers with (black bars) and wit hout (grey bars) their young infants (<3 months old) to troop-mate s during the fruiting season. Mothers carrying their infants (black bars) stayed significantly closer to other females carrying young infants ( P = 0.039) and to males ( P = 0.007) than when the mothers were without their infants (gre y bars). The physical presence of an infant, however, did not affect a mother s proximity to females without babies. 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 FruitingWetDryTransition 0 5 10 15 20 25 30 35 40 45 50 MalesFemales w/ infantFemales* P = 0.007 P = 0.039 Proximity score (X SE) X groom rate (min/h) SE
167 Figure 4-11. Grooming durat ion (%) between females with infants and adult troopmates. Black bars represent the % of grooming that females with infants (n = 4) received from other females with infants ( n = 3), females (without infants) ( n = 3), and males ( n = 3). Grey bars represent the % of grooming that females with infants groomed females with infants, females (without infants), and males. Females with infants devoted the most grooming to other females with infants, while females without in fants received the least. All of the grooming that adult males received from females with infants was directed towards the alpha male. Females with infants received the majority of adult dyadic grooming from females (both with and wit hout infants). 0 5 10 15 20 25 30 35 40 45 Female w/ infantFemaleMale % of grooming duration
168 Figure 4-12. Total duration of time (min) that mothers with (black bars) and without (grey bars) their infants (< 3 mont hs old) groomed and received grooming from troop-mates during the fr uiting season. Mothers ( n = 4) both groomed and received grooming significantly mo re when they were carrying their infants (black bars) than when they were without their inf ants (grey bars) ( P < 0.001 and P < 0.001, respectively). The diff erence was not due to a disparity in opportunity, however, as there was no statistical difference in the number of scans that the mothers were observed with ( n = 123) and without ( n = 101) their infants during this time ( P > 0.10). 0 50 100 150 200 250 GroomReceiveGrooming duration (min) P <0.001
169 CHAPTER 5 DO FEMALE BROWN CAPUCHIN MON KEYS US E AFFILIATIVE BEHAVIOR TO MEDIATE STRESS? Introduction This chapt er targets the poor ly understood nexus that lin ks the dynamics of female stress response (defined as the increase in the stress hormone cortisol) with social behavior and relationships. The glucocorticoid (GC) cortisol, a stress hormone, is commonly measured as an indicator of stress. Upon encountering a stressor (i.e., any extrinsic or intrinsic threat to homeostasis) GCs are released into the bloodstream in response to activation of the hypothalamus-pit uitary-adrenal (HPA) axis (Wingfield et al. 1997). Over the short term, acute stre ss response is an adaptive physiological condition, which, upon encountering a threat, regulates the diversion of energy from long-term storage to meet short-term needs, t hus increasing prospects of immediate survival. Over extended periods, however, t he metabolic costs of sustained stress rise steeply (Sapolsky et al. 2000), adversely affe cting female health and reproductive ability, fetal development and the developm ent of offspring social skills, and stress-coping abilities (Sachser et al. 1998; Ba rdi et al. 2005; Sapolsky 2005; Shively et al. 2005; Wadhwa 2005). Therefore, in order to protect current and future reproductive success, a female should take an ac tive role in managing stress. Both social (Sapolsky 2005; Soto-Gamboa et al. 2005) and environmental factors contribute to physiological stress in captivity and the wild (Boinski et al. 1999b; Lynch et al. 2002 ; Pride 2005a,b). While field studies tend to focus on the variables correlated with stress levels of wild animals (Holekamp and Smale 1998; Foley et al. 2001; Lynch et al. 2002; Rogovin et al. 2003; Altmann et al 2004; Weingrill et al 2004; Bales et al. 2005; Beehner et al. 2005; Creel 2005; Goul d et al. 2005; Pride 2005a,b), experimental
170 manipulations of captive anima ls assess the importance of so cial relationships in the management of stress among a few taxa, specifically and most commonly rodents and non-human primates (Tamashiro et al. 2005). Thus, there is a paucit y of data directly investigating how socioecological stress co-v aries with social relationships in wild populations. Stress and Sociality The social environment of group-living species is a balance of deleterious consequences and adaptive benefits. Specifically in regard to social relations hips, individuals may face negative physiological e ffects as a result o f, for example, their dominance status (Chapter 3) but be provided with a (possible) means to reduce neuroendocrine response to stressful situati ons: In some animals (e.g., guinea pigs (Sachser et al. 1998; Kaiser et al. 2003; He nnessy et al. 2006); prairie voles (Carter et al. 1995); tree shrews (von Holst 1998); squi rrel monkeys (Saltzman et al. 1991); marmosets (Norcross and Newman 1999; Ru kstalis and French 2005); and rhesus macaques (Gust et al. 1994), including humans (K irschbaum et al. 1995; Heinrichs et al. 2003; Rosal et al. 2004), the presence of an adult partner provides stress reducing effects. Often referred to as social support (Sachser et al 1998), the ability to reduce cortisol reactivity to a st ressor cannot be provided by any conspecific or group-mate but rather is limited to bonding partn ers (Mendoza et al. 1991; Cart er et al. 1995; Sachser et al. 1998; von Holtz 1998; Rukstalis and Fr ench 2005; Hennessey et al. 2006; but see Saltzman et al. 1991; Ka iser et al. 2003). While the importance of bondi ng partners is most studied in monogamous species (i.e., male-female pair bonds), the adaptive functions of social bonds are likely provided both interand intra-sexually and also within other types of social systems. The
171 restriction appears to be not the type of social system itself, but rather the affinity of individuals to form strong within-group partner preferences (Hennessy et al. 2008). For example, plasma cortisol le vels of male guinea pigs (Cavia aperea porcellus ), a species with a single male, multi-female social system similarly increased subsequent to being removed from their social colony and placed in a novel environment ei ther in isolation, with an unfamiliar female, or with a familia r but unbonded female. When placed with a female to whom they were bonded, however, there was a significant decrease in the males cortisol reactivity (Sachser et al 1998). Similar results were found when testing the effects of familiar unbonded versus bonded ma le partners (Kaiser et al. 2003) and bonded versus unfamiliar male partners (Hennessy et al. 2008) on the cortisol response to a novel environment of female guinea pigs. Further, while a harems female guinea pigs naturally space themselves out and displa y minimal interaction, there was evidence that captive female-female pairs may demons trate HPA buffering indicating the ability to, given opportunity, form intrasexual social bonds (Hennessy et al. 2008). Females are more prone than males to rely on their social network in times of stress, a difference so striki ng that it is considered to be the primary gender difference in behavioral response to stress, especially among humans. Further, women are not only more likely than men to choose to af filiate during stress, but women choose to affiliate with other women (Taylor et al 2000). Taylor and colleagues proposed that while the fight-or-flight stress response is adaptive to males, such a response would not address the challenges faced by females; a mo thers decision to eith er fight or flight would similarly endanger her offspring, oft en rendering them alone and unprotected. As such, the tend-and-befriend pattern evolved in group-living females to protect both the
172 female and her offspring. Empirical ev idence supporting the adaptive significance of female social bonds is provided by a wild population of savannah baboons ( Papio cynocephalus ); females that are more socially integrated within their groups had higher rates of reproductive success (infant surviv al to year one) than other females, regardless of rank (Silk et al. 2003). Fu rther highlighting the importance of female social networks, female chacma baboons ( Papio hamadryas ursinus ) experienced elevated GC levels when their social bonds were disrupted or became unstable (Engh et al. 2006a) and expanded their social networks when a preferred par tner was lost to predation (Engh et al. 2006b). The presence of non-human primate soci al bonds is most commonly quantified through measures of grooming and time spent in proximity. T he physiological benefits of grooming have been demonstrat ed empirically in several studies involving captive non-human primates. For example, Boccia et al. (1989) and Aureli et al. (1999) showed that in a female pigtail macaque ( Macaca nemestrina ) and female rhesus macaques ( Macaca mulatta ), respectively, the receipt of groom ing lowered heart rate. Similarly, Gust et al. (1993) provided evidence fo r a negative correlation between affiliative behaviors (including proximity and the receipt of grooming) and cortisol levels in female rhesus macaques. Such evidence lends to the notion that grooming is an effective social outlet for reducing stress (Abbott et al. 2003) and is one of the primary coping mechanisms used by female non-human primates to reduce HPA axis activity (Wittig et al. 2008). The mechanism for such likely st ems from the attachment-caregiving system between a mother and infant, and thus is (at l east in part) oxytocin-mediated (Taylor et al. 2000). In general, oxytocin both inhibits the release of GCs (Neumann et al. 2000b;
173 Legros 2001) and promotes affiliative behav ior (Martel et al. 1993; McCarthy and Altemus 1997; Carter 1998; Uvnas-Moberg 1998; Insel and Young 2001). The degree to which oxytocin is involved in HPA ax is suppression, however, is dependent on both species and reproductive state (Neum ann et al. 2000a; Brunton et al. 2008). Stress Hormones and Female Reproductive Condition Pregnancy and lactation Female stress hormone concentration is affected by reproductive condition (Brunton et al. 2008). For example, am ong baboons, female cortisol levels were signific antly higher during pregna ncy than during other reproducti ve stages (Weingrill et al. 2004; Gesquiere et al. 2008), a pattern re plicated by several other primate taxa including tamarins (Ziegler et al. 1995; Bales et al 2005), marmosets (Smith and French 1997), lemurs (Cavigelli 1999), and humans (Lockwood et al 1996). During late pregnancy, especially peripartum, female human and non-human prim ates experience increased basal GC levels. Such hypercortisol ism, however, is not to be viewed as an internal stressor but rather a necessary adaptive condition of gestation and labor, serving to support fetal growth and develop ment and link fetal organ maturation with parturition (Pepe and Albrecht 1995). This incr ease in GC concentrations is due to the production of placental corticotropin-releasing hormone (CRH), which increases overall GC concentration (Goland et al. 1994; Lockwood et al. 1996). Despite the primate pattern of elevated basal GC levels during late pregnancy, it has been shown in various non-human mammals (with research heavily biased towards the rat) that the responsiveness of the HPA axis to physical and psychological stressors decreases during late-pregnancy and is ma intained through lactation (Cook 1997; Windle et al. 1997; Shanks et al. 1999; Brunton and Russell 2003; Tu et al. 2005;
174 Tilbrook et al. 2006). For example, Tilb rook and colleagues (2006) show that, in response to isolation and restraint stress, cortisol levels increased significantly in non-lactating ewes but were not affected in lactating ewes. Additionally, the physical presence of their suckling infants further dec reased the mothers cortisol concentrations as compared to lactating ewes with lambs absent and to non-lactating females. In response to physical stress, a similar pattern was found for human females (Altemus et al. 1995; Kammerer et al. 2002). In respons e to psychosocial stress (e.g., public speaking), however, the physiological state of lactation did not appear to provide an overall attenuating effect on the human HPA axis, whereas t he short-term act of nursing did result in temporary endocri ne suppression (Altemus et al. 2001; Heinrichs et al. 2001). The adaptive benefits of attenuated maternal stress response are well-known, namely to prevent excessive levels of ci rculating GCs that could impair healthy development of offspring, proper maternal care, adequate and nutritious milk supply, and the physical (e.g., immune function) and psychological well-being of the mother (Altemus et al. 1995; McCormick et al. 1995; Vallee et al. 1997; Weinstock 2001). For example, pregnant and lactati ng women reported feelings of increased calmness and reduced anxiety and moodiness (Carter et al. 2001; Heinrichs et al. 2001; Glynn et al. 2004), and the abuse of infants by rhesus macaque mothers was linked to elevated HPA axis activity (Maestripieri et al. 2005). Reduced HPA axis activity of the maternal brain, therefore, is a protective mechanism by which reproductive success is improved and mothers are better able to mentally and em otionally cope with hormonal fluctuations (Slattery and Neumann 2008).
175 The mechanisms underlying suppression of maternal stress-responsivity, however, are yet to be fully understood. Although oxytocin (necessary to stimulate milk let-down and maternal behavior) suppresses GC production, the release of oxytocin during parturition and in response to suckling did not explain the attenuation of stress responsivity observed in pregnant and lactating ra ts. Rather, oxytocin was an inhibitor of the HPA axis only in virgin females (Neumann et al. 2000a). HPA axis hyporesponsiveness during pregnancy appears to be due to endogenous opioids, whereas low estrogen levels may contribute to reduced GC response during lactation. An alternative, and possibly complementary, explanation for lactation-induced HPA axis hyporesponsiveness may be the neuropeptide prolactin whic h, like oxytocin, is necessary for milk production and is released in response to suckling (Brunton et al. 2008). In virgin rats, prolactin infusion (which mimics peripartum physiology) reduced anxiety-related behavior and HPA axis activity (Torner et al. 2001; Donner et al. 2007), and prolactin was more effective than oxytocin in reducing cortisol secretion in response to stress in both lactating and non-lactating sheep (Cook 1997). Estrus cycling Considerably less has been publis hed regar ding the relationship between stress hormones and the female ovarian cycle of es trus and anestrous, likely because of the difficulties in studying such transient c onditions. Weingrill and colleagues (2004) reported that, overall, cycling female baboon s had lower cortisol values than pregnant and lactating females, a pattern also found among cotton-top tamarins (Ziegler et al. 1995). Because estrogen can stim ulate the HPA axis therefor e resulting in increased production of GCs (Coe et al. 1986; Altemus et al. 1995), elevated cortisol levels are expected during the periovulatory phase (i.e ., the estrus period) when estrogen levels
176 are high (Weingrill et al. 2004) Support for this prediction comes from studies on the cotton-top tamarin (Ziegler et al. 1995), common marmoset (Ziegler and Sousa 2002), and dairy cow (Lyimo et al. 2000). On the ot her hand, several studies provide empirical evidence that do not support this prediction. For example, a study of dairy cattle showed no change in cortisol levels between estrus and anestrous females (Walker et al. 2008). Among wild chacma baboons ( Papio hamadryas ursinus ), there were no significant differences in periovulatory cortisol levels as compared to other portions of the female cycle, but the aut hors acknowledged that the limited number of samples collected during the various phases may have precluded detection of actual patterns (Weingrill et al. 2004). Lahoz and colleagues ( 2007) reported similar findings for captive Cebus : Despite a threefold increase in periovula tory estradiol concentrations, circulating cortisol levels remained unchanged throughout the female cycle. Hypotheses Using brown capuchins ( Cebus apella ) as my study system, I investigated the relationship between female cortisol levels, social relationships and reproductive conditio n, specifically addressing if and how females use affiliative behavior to mediate stress. My study of st ress response and management is the first to examine the hormonal correlates of social behavior for female brown capuchins in the wild. Although female C. apella are widely accepted to be a fema le-bonded species, with females dedicating more affiliative behavior (grooming and close proximity) to females rather than to males (Di Bitetti 1997), the results of my study (Chapter 4) indicate deviation from that pattern. Female-fem ale relationships were most evident in the presence of a young infant, and females were generally more closely bonded to males than to females. Further, temporal va riation in a females social network, rather than stable,
177 long-term relationships, appears to be the norm fo r this population. Do the shifts in female social networks co-vary with ecolog ical, social and reproductive stressors? While controlled laboratory studies indicate a reciprocal relationship between stress and social support, is such a dynamic empiricall y evident in a wild population? To address these questions, I studied the following hypotheses: Female cortisol levels are inversely associated with rates of affiliative behavior. Female cortisol levels vary with monthly variations in affiliation. Female cortisol levels va ry with reproductive condition. Female reproductive condition affects affiliative behavior. Methodology Site Description Raleighvallen (RV), a 7812 km2 reserve consisting of primary tropical rain forest, is located within the 1.6 million ha Central Suri name Nature Reserve in Suriname. The main study site measured approximately 2 km2 and was covered by an extensive trail system. RVs flora and fauna are effectively undisturbed in the historical period, and an intact array of potential predators and competitors are present (Reichart 1993). The soils of the Guyanan Shield (including French Guiana, Suriname and Guyana) are nutrient-poor and highly weathered as they are derived from Precambrian bedrock. Fleshy fruits are low in abundance and available fruit resources typically occur in small patches (< 5 m diameter) (Boin ski et al. 2002). Additionally, much of RV is composed of expansive bamboo patches and dense liana forest t hat restricts visibilit y (Boinski et al. 2003).
178 Study Animal The brown capuchin monkey ( Cebus apella ) is one of the four traditionally recognized species in the genus Cebus (Hill 1960). Capuchins li v e in multi-male, multifemale polygamous groups that are typically female philopatric. Adult males and females maintain separate dominance hierar chies; in general, males are dominant to females. C. apella social groups are typically compri sed of 12 to 27 individuals, with an average troop size of 17 members. Troop membership and relationships remain relatively stable, except for subadult male s who emigrate from their natal group and may subsequently transfer several times. Brown capuchins reside in a variety of habitats encompassing a large geographic range from Columbia to Argentina (Fragaszy et al. 2004). C. apella are omnivores; the bulk of t heir diet consists of fruit supplemented with seeds, vegetation, arth ropods, and vertebrates (Terborgh 1983). Female brown capuchins demonstrate a cons picuous behavioral estrus (Janson 1984; Carosi et al. 1999) that is reliably indicative of the periovulatory phase; the length of the female cycle for this species is, on aver age, 20.6 1.6 days (Carosi et al. 1999). The study troop (Troop A) was one of four brown capuchin groups commonly observed within the study site; the primary study subjects were the adult females and their relationships to the adult members of the group. For most of the study, Troop A consisted of 27 individuals, seven of which were adult females. The seven females of focus encompassed the spectrum of r eproductive conditions: cycling, pregnancy (primiparous and multiparous), and lactating. During the study, one pregnant female disappeared; her disappearance coincided with her predicted du e date, and thus it is possible her (presumed) death was related to pregnancyor labor-related complicated (Table 5-1). Relatedness among troop member s is unknown; however, genetic analysis
179 of the brown capuchin population in RV is currently underway. The study troop had been habituated to human observers since 1998, and all individuals were recognized based on body size, color patterns, and other i dentifying characteristics such as scars, moles, and ear shape. Throughout the study period, inter-observer reliability tests were conducted monthly to ensure consistent and accurate identification of all troop members. Data Collection From January to December 2006, I and/or field assistants followed Troop A for up to 10 h each day (beginning at sunrise). Th e group was followed, on average, 28 days per month, composing a total of 2284 h of observation (not incl uding out of view time). Female proximity data (within a 4 m radius ) were collec ted via individual instantaneous scans ( n = 787) and converted to quantified prox imity scores, abbreviated as MPS (proximity to adult males) and TFPS (proximity to adult females). A detailed description of proximity data collection and proximity score calculations can be found in Chapter 4. Grooming and reproductive behavior were noted ad libitum (Altmann 1974). Behavioral and Reproductive Definitions I defined a female bond as one in which a fe males primary adult social partner (in terms of grooming and/or proximity patte rns) was another female. An adult dyad was considered to have a strong bond (based on grooming or proximit y) if their dyadic grooming duration or proximity score fell wit hin the top 10% for t he group (Silk et al. 2006b). Grooming bout duration and frequency was recorded for all observed occurrences. A grooming bout involv ed one individual gr ooming another, and concluded when the grooming ended or was interrupt ed by another behavior for more than five seconds. A
180 new grooming bout began with a new dyad, or whenever the grooming continued or changed direction within the same dyad. Only grooming bouts involving at least one female (as actor or recipient), and where both participants could be i dentified (at least to the age and sex class) were included for analysis. All grooming bouts were counted as dyads; thus, if two monkeys were grooming the same individual concurrently, the episode was scored as two separate actorrecipient grooming bouts. Grooming bouts involving females in estrus (as actors or re ceivers) were included in analysis only when considering the dynamic between affiliati ve behavior and reproductive condition. The focal females encompassed three repro ductive conditions: cycling (females in any stage of the estrous cycle), pregnant (assigned post hoc from the birth of an infant and starting at the fi nal detumescence of the estr us cycle), and lactating (the period following birth until the resumption of cycling) (Crockford et al. 2008, p. 257). Lactation and the resumption of cycling followi ng lactation were not mutually exclusive, however, as females tended to resume (at le ast behavioral) estrus while still nursing; mothers resumed estrus-like cycling when their infants ranged in age from 5 to 10 months. Hormonally, I classified these female s as lactating rather than cycling because their cycles were not yet regular and they nursed on a regular basis. Behaviorally, when examining the relationship between reproductive condition and affiliation, I classified these females as being in es trus when they displayed species-typical proceptive behavior. Based on cyclical pa tterns of behavioral proceptivity common to C. apella (e.g., eyebrow raising, grinning, wheez vocalizations, head cocking, chest rubbing, touching (a male) and then running away, backing into the lap of a male, mounting) (Janson 1984; Carosi et al. 1999; Ehmke, personal observation), I was able
181 to ascertain when a cycling female was in estrous. For purposes of my study, I classified a cycling female as anestrous if she was not in estrus. Hormonal Sampling and Analysis Fecal samples were collected opportunistica lly, with the goal of collecting at least one sample per femal e ( n = 7) per 15-day interval (Engh et al. 2006a,b; Lynch et al. 2002). Fecal samples were collected only if the female was observed defecating and her identity ascertained with certainty. U pon collection, each sample was immediately placed in an empty plastic vial and stored inside a thermos with an instant chemical cold pack. Samples were processed within th ree hours of collectio n and registered, on average, at 44F. The samples were proc essed using Solid Phase Extraction (SPE) according to Ziegler and Wittwer (2005): 0.1 g of fecal materi al was mixed with 2.5 ml of distilled water and 2.5 ml of ethanol. The mixture was then hand-shaken for 5 min and centrifuged for 10 min. Two ml of the s upernatant was removed and passed through an Alltech Prevail C18 Maxi-Clean SPE Cartridge (Alltech, Deerfield, IL) and stored. Samples were shipped to The Wisconsin Regional Primate Res earch Center for analysis. Using Enzyme ImmunoAssay, eac h sample was analyzed for cortisol (F) concentration (ng/g); the assays were va lidated for accuracy and parallelism using internal controls. A total of 265 fecal samples were colle cted from Troop A females. A baseline fecal cortisol level for each female was established by calculating the mean after removing the highest and lowest 5% of val ues (Peel et al. 2005). These baseline cortisol values ( n = 237) were used when analyzing the relationship between cortisol and reproductive state. When ex amining the relationship between female cortisol levels and affiliation, however, all cortisol values ( n = 265) were included in analysis.
182 Statistical Analysis When feasible, data were analyz ed using para metric statistical tests; when data failed to meet the assumption of normality, however, non-parametric alternatives were used. When comparing two independent samples (e.g., groom rates of pregnant versus lactating females), I used the independent sa mples t-test/Mann-Whitney U test; when comparing two related samples (e.g., mean rates at which pregnant females groomed males versus females), I used the paired sa mples t-test/Wilcoxin signed-rank test. When comparing more than two samples, I used either one-way ANOVA in conjunction with Tukeys post hoc tests or the Kruskal-Wallis one-way ANOVA in conjunction with Mann-Whitney U tests as the nonparametric alternative to post hoc tests. Correlations between parameters were evalua ted using either Spearmans rs (Excels Analyse-It statistical package) or, to cont rol for a third variable, partial correlation. Analyses were performed using SPSS 11.0; the significance le vel for all tests (two-tailed) was set at 0.05. Mean values are pres ented with standard error values. Results Association between Social Proxim ity and Female Cortisol Levels Overall, there was no troop-wi de association between females mean proximity scores and cortisol levels (TPS: rs = -0.43, P = 0.34; TFPS: rs = -0.64, P = 0.12; MPS: rs = 0.36, P = 0.43) (Table 5-2). Similarly, t here were no associations between an individual females cortisol levels and monthly proximity to males ( rs = -0.31 to 0.58, P = 0.10 to 0.73) or to females ( rs = -0.06 to 0.42, P = 0.22 to 0.88) (Table 5-3). While most females showed a positive (although no t necessarily significant) relationship between proximity scores and time spent in a central position within the troop (Table 5-4), there were no associations between central position and cortisol levels
183 ( rs = -0.12 to 0.53, P = 0.11 to 0.90). In regards to monthly prox imity strong bonds (the top 10% of females monthly proximity scores), there were no individual associations between months with versus without such st rong bonds and cortisol levels. Further, data did not indicate any directional descrip tive trend in female cortisol levels: Two females had higher mean cortisol levels in months with proximity strong bonds as compared to those months without proximity strong bonds; two females had comparably higher mean cortisol levels in months wit hout proximity strong bonds; and the mean cortisol levels of two females were nearly equivalent in mont hs with and without proximity strong bonds (Table 5-5). Association between Grooming (gi ven) and Female Cortisol Levels Overall, there was no troopwide association between female cortisol levels and the monthly rates at whic h females groomed females ( rs = -0.36, P = 0.43). The rate at which females groomed males, however, was significantly associated with female cortisol levels ( rs = -0.77, P = 0.045). While there was a strong correlation between a females rank and the rate at which she groomed males (r = 0.78, P = 0.04), there was no association between female rank and overall cortisol level ( r = 0.39, P = 0.38). Further, after controlling for rank, the relations hip between female cortisol levels and the grooming of males remains st rong and nearly significant ( r = -0.80, P = 0.058). Sample sizes (i.e., grooming rates) were too small and cortisol levels too variable to test statistically for monthly associations between the cortisol le vels of an individual female and the rate at which she groomed females and males. For two females, Jane and Carol, however, there was a significant, positive correlation between the number of adult troop-mates they groomed in a month and their monthly cortisol concentrations ( rs = 0.823, P = 0.01; rs = 0.748, P = 0.02, respectively); for t he remaining five females,
184 no association existed between monthly groo ming network size and cortisol values ( P = 0.23 to 0.92). In regards to monthly grooming strong bonds (the top 10% of adult dyadic grooming durations given by females) and cortisol levels, data did not indicate any directional descriptive trends: Two females had higher mean cortisol levels in months with grooming strong bonds as com pared to those months without grooming strong bonds; three females had comparably higher mean cortisol levels in months without grooming strong bonds; and the mean cortisol levels of one female were nearly equivalent in months with and without grooming strong bonds (Table 5-6). Association between Grooming (recei ved) and Female Cortisol Levels Overall, there was no troopwide association between female cortisol levels and the monthly rates at which females received grooming from females ( rs = -0.57, P = 0.18) or from males ( rs = -0.25, P = 0.59). Sample sizes (i.e., grooming rates) were too small and cortisol levels too variable to test statistically for monthly associations between the cortisol levels of an individual female and the rate at which she received grooming from females and males. However, there existed no correlation between the number of adult troop-mates that groomed a female eac h month and her monthly cortisol values ( P = 0.15 to 0.84). In regards to monthly gr ooming strong bonds (the top 10% of adult dyadic grooming durations re ceived by females), there resulted no significant associations between the pres ence of monthly grooming strong bonds and cortisol levels for any female. However, a descriptive trend emerged: All females (with the exception of the most s ubordinate female) had higher cortisol levels in the months characterized by the presence of str ong grooming bonds (grooming received) as compared to the months wit hout such bonds (Table 5-7).
185 Reproductive Condition, Female Cortis ol Levels an d Affiliative Behavior Cortisol levels Comparing all females in various states of reproductive condition (anestrous, estrus, pregnancy, and lactation), ther e was a troop-wide associat ion between reproductive condition and female cortisol levels ( P = 0.013). Controlling for season, this association remained nearly significant ( P = 0.067). Overall, pregnant females ( n = 2) had the highest cortisol levels (mean SE = 75.29 17.01), while lactating females ( n = 4) had the lowest cortisol levels (mean SE = 43.26 4.90). Posthoc analysis revealed that lactating females had significantly lower cortisol levels than anestrous ( P = 0.027), estrus ( P = 0.018) and pregnant ( P = 0.033) females (Figure 5-1). Due to limited cortisol samples and variable cortisol data, it was not possible to determine statistically whether there existed interor intra-individual differences in cortisol levels by reproductive condition. Pregnant females were represented only by Jane (dominance hierarchy ranking 3 of 7) and Tinkerbell (ranking 7 of 7), the two females with the highest baseline cortisol levels (Chapter 3). While I cannot rule out the possibility that these individuals were naturally more prone than other females to cortis ol reactivity of extr insic variables, their pregnant condition appears to have been a conf ounding factor. Tinkerbells cortisol levels increased steadily throughout her pregnancy and were strongly and positively correlated with gestation length ( rs = 0.94, P = 0.005) (Figure 5-2). Janes cortisol levels, on the other hand, were not associated with gestation length ( rs = 0.10, P = 0.715), but in her final month of pregnancy (also the month of her presumed death), Janes mean cortisol level was significantly higher than in all other months in which she was pregnant (P = 0.01 to 0.03) (Figure 5-3).
186 Lactating females ( n = 4) represented the span of the dominance hierarchy (ranks 1, 2, 5, and 6 of 7) and thus pr ovided a natural (troop-level) co ntrol for the effect of rank on lactating female cortisol levels. Cortisol levels of lactating females were associated with incremental age groupings of their infants ( P = 0.005), although in no directional pattern (i.e., the cortisol le vels of lactating females did not consistently increase (or decrease) over time as their infants aged and lactation demand lessened). Cortisol levels were highest (mean SE = 68.72 17.26) for lactating females when their infants were 3 months of age and lowest when their in fants were 6 to 9 months of age (mean SE = 24.34 3.06) (Figure 5-4). These time spans, however, generally corresponded with the fruiting season (February to April) and dry season (August to October), respectively, and thus introduces seasonality as a confounding variable. Controlling for season, there existed no correlation between fe male cortisol levels and lactation over time (3-month increments) ( r = -0.11, P = 0.19). Proximity In terms of proxim ity, lactating fema les (i.e., those with dependent infants) had significantly higher proximity scores with adult females (TFPS) (mean SE = 29.02 6.05) than did cycling (mean SE = 1.94 0.70) (P < 0.0001) and pregnant females (mean SE = 3.50 2.18) ( P = 0.001). The pattern holds tr ue for the proximity to adult males (MPS) of lactating (mean SE = 26. 88 4.93), cycling (mean SE = 18.11 5.18), and pregnant (mean SE = 7.14 4.23) females, although the difference is significant only when the MPS of lactating females is com pared to the MPS of pregnant females ( P = 0.023). Further, there was a negative correlation ( rs = -0.44, P = 0.005) between infant age and the TFPS of mothers ( n = 4), but no relationship between infant age and mothers MPS ( rs = -0.06, P = 0.74) (Table 5-8). Ov erall, there were no
187 statistical differences between TFPS and MPS at any 3-month incremental infant age categories ( P = 0.09 to 0.59), but it is interesting to note that for mo thers with dependent infants, TFPS remained higher t han MPS generally for the first 6 months of their infants lives, after which MPS was higher than TFPS (Table 5-8); individually speaking, this pattern held true for two of the mothers, while the re maining two mothers had TFPS values that were higher than MPS values for only the first 3 months of their infants lives. For pregnant females ( n = 2), there was no associat ion between gestation length and proximity to females ( rs = -0.09, P = 0.81) or males ( rs = 0.24, P = 0.51). The pregnant females maintained a higher (nonsignificant) MPS than TFPS during each month of pregnancy (Table 5-9). As proximity data were not collected on estrus females, I was unable to test for the effect of estrus cycli ng on female proximity scores. Estrus females, however, overwhelmingly d edicated their close attention to following males, and specifically the alpha male. While in anestrous, cycling females (n = 3) had significantly higher MPS (18.11 5.18) than TFPS (1 .94 0.70) (P = 0.001). Grooming (given) In terms of grooming behavio r, reproductive condition did not have a significant effect on the rates at which females groomed females ( P = 0.293) or males ( P = 0.196). Lactating females groomed both females and males at the highest rates (0.0201 0.0057 min/h, 0.0357 0.0062 min/h, respecti vely), while pregnant females least groomed both females and males (0.0085 0.0058 min/h, 0. 0173 0.0088 min/h, respectively). In each reproductive c ondition, females groomed males at a higher (non-significant) rate than they groomed females (Figur e 5-5). The effect of infant age on the rates at which mothers groomed males approached significance ( P = 0.061), with lactating females most grooming males when their infants
188 were 0 to 3 months old (0.0528 0.0132 min/h) and 9 to 12 months old (0.0551 0.0145 min/h) as compared to when the inf ants were 3 to 6 and 6 to 9 months old (mean SE = 0.0212 0.0096); infant age did not affect the rate at which mothers groomed females ( P = 0.249), females groomed mothers ( P = 0.107), or males groomed mothers ( P = 0.420). The rate at which mot hers groomed females, however, decreased steadily with infant age, and the two variables were modestly correlated ( rs = -0.37, P = 0.021) (Figure 5-6). A strong, positive correlation occurred between gestation length and the rate at which Jane, one of the pregnant females, groomed females ( rs = 0.95, P = 0.05), but there was no such association for t he other pregnant female, Tinkerbell ( rs = 0.22, P = 0.72). Further, in her fi nal month of pregnancy (Octobe r), Jane intensified her grooming effort towards the dominant male and female. In October, Jane dedicated 48% of her grooming duration to the alpha ma le, whereas in all previous months she dedicated, on average, 35% of her (non-estrus) grooming to him. Si milarly, in October, Jane devoted 45% of her gr ooming duration to the dom inant female, whereas previously, this grooming investment averaged 11%. Tinkerbells final month of pregnancy occurred after the conclusion of t he study, and so I was unable to determine if her peripartum grooming behavi or followed a similar pattern. Cycling females groomed males at nearly two times the rate that they groomed females (Figure 5-5). When in estrus, cyc ling females devoted 97.9% of their grooming behavior towards the alpha male, whereas during anestrous, only 20.5% of their grooming was directed towards the alpha male.
189 Grooming (received) There was a significant effect on the rates at which females of different reproductive condition were groomed by females ( P = 0.010); this difference was attributable to the low rate of grooming t hat the pregnant females received from other females (0.0004 0.0004 min/h). Both lactating and cycling (estrus and anestrous periods combined) females received signi fic antly more grooming from females (0.0285 0.0104 min/h, 0.0145 0.0067 min/h, respectively) than did pregnant females ( P = 0.003, P = 0.049, respectively), while there was no difference between the rates at which females groomed lactating and cycling females ( P = 0.209). Further, only lactating females received more grooming fr om females than males, and this difference was significant ( P = 0.012) (Figure 5-5). There was a nearly significant difference in the rates at which males groomed females of different re productive condition ( P = 0.056), with males focusing their grooming behavior on cycling (estrus and anestrous periods combined) females (0.0188 0.0076 min/h) as compared to pregnant (0.0013 0.0010 min/h) and lactating (0.0060 0.0025 min/h) females (Figure 5-5) When in estrus, cycling females were groomed only by the alpha male; neither subordinate male nor any females were observed to groom estrus females. Discussion Social Affiliation and Cortisol Levels Overall, group-level data did not support the hypothesis that female cortisol levels are inversely associated with affiliative behavi or. Specifically, there was no discernable relationship between fecal GC concentration and proximity to adult male and/or female troop-mates, the receipt of grooming from males and/or fe males, or the grooming of
190 females. There was, however, an inverse a ssociation between female cortisol levels and the rate at which females groomed male s, a surprising finding given the evidence that portrays the physiological benefit of the receipt of grooming (Boccia et al. 1989; Gust et al. 1993; Aureli et al. 1999). This unexpected resu lt supports the findings of Shutt et al. (2007), who found tha t, in female Barbary macaques ( Macaca sylvanus ), it was the giving rather than the receiving of grooming that was associated with reduced stress levels. In the earlier captive studies, the researchers measured the monkeys immediate physiological response to grooming (v ia blood collection or heart monitoring implants); like Shutt et al. (2007), howev er, the current st udy was done on a wild population and the physiological response to grooming behavior was measured via fecal (i.e., pooled) GC concent rations. Therefore, the short-term physiological benefits of the receipt of grooming do not appear to lowe r stress levels in the longer term (Shutt et al. 2007). Why the link between the grooming of ma les (specifically) and lower female cortisol levels? Female rank may be a c onfounding factor if more dominant females, who have greater access to males, generally hav e lower stress levels. This was not the case, however, as there was no correlation between female social rank and baseline cortisol levels (Chapter 3), and after contro lling for rank the dynamic remained between female GC concentration and the grooming of ma les. Alternatively, reduced HPA axis activity may have resulted from the expect ation of a strengthened social relationship, supporting the notion that grooming serves to initiate, maintain and/or strengthen social bonds (Stammbach and Kummer 1982; Dunbar 1996; Lehmann et al. 2007; Fedurek and Dunbar 2009). That the grooming of female troop-mates was not found to be
191 associated with a decrease in female cortisol le vels indicates that grooming it itself is not an effective social outlet for reducing stress, but may be dependi ng upon the partner involved and the potential benefit s of strengthening that particular dyadic relationship. This finding further supports the conclusions that females in a female-philopatric species may not necessarily view other female s as their most benefic ial social partners, and that female brown capuchins in RV are generally more bonded to males than to females (Chapter 4). It is interesting to note, however, that despite the possible long-term physiological benefits of grooming males, females rarely (if ever) were observed to compete for groom ing access to males. Data also did not support the individual-leve l hypothesis that female cortisol levels vary with monthly variation in affiliative rela tionships. Such a short-term association, however, may have been lost due to variability in GC concentrations and in low rates of social behavior. When com paring the months in which females had strong bonds (the top 10% of dyadic interaction) against less sociable months, however, (non-significant) patterns began to emerge, especially in regard to the receipt of grooming: Female cortisol levels were higher in months in which they demonstrated at least one strong bond via the receipt of grooming as compared to months in which they had no strong bonds. This positive relationship may appear puzzling given the relaxing and stressreducing benefits provided by the tactile stimulation of being groomed. As discussed above, however, such physiological benefits ma y be so temporary that longer-term fecal GC concentrations are unaffected. Alt hough I cannot, with any degree of certainty, attribute a causal link between reduced cortis ol levels and strong grooming bonds, it is possible that this relationship indirectly supports the idea that social support (i.e.,
192 specific or particularly strong bonds) is e ssential for mediating stress (Mendoza et al. 1991; Carter et al. 1995; Sachser et al 1998; von Holtz 1998; Rukstalis and French 2005; Hennessey et al. 2006; Wittig et al. 20 08), such that females may have received more grooming from specific bonding partners because they were more stressed. Each females grooming network typically incl uded one to three adult conspecifics (male and/or female) that categor ically produced multiple monthly strong bonds, while Tinkerbell, the female with the highest mean GC concentra tion, did not have a single adult grooming partner that produced multiple monthly strong bonds. Thus, it may have been Tinkerbells lack of adult social s upport that hindered effective stress management. Alternatively, the positive association between female stress levels and their receipt of grooming may have been confounded by an outside variable, such as the presence of a young infant. The cortisol leve ls of lactating females were highest when their infants were less than th ree months of age, and this time span corresponds to the period of greatly intensified grooming effort directed towards mothers (Chapter 4). The increase in GC levels of lactating fema les may have resulted from the unsolicited attention they and their babies received and the lack of control they were able to exert over their grooming part ners (Crockford et al. 2008) obscuring the HPA axis suppressive effect of lactation. Overall, however, lactating females had the lowest cortisol values (as compared to pr egnant and cycling females) (see below), and although lactating females demonstrated mu ltiple strong bonds via the receipt of grooming in the month of February (when th ey each had an infant that was less than three months old), the females without infants also demonstrated strong bonds via the
193 receipt of grooming during the month of F ebruary (Chapter 4). Therefore, infant presence alone does not account for the dy namic between the increased receipt of grooming and female cortisol levels. Aside from February, female s most demonstrated strong bonds through the receipt of grooming in the months of (in descending order) May, September and April, an ex panse of time that controls for the effect of season and reduces the impact of infant presence on t he relationship between stress and social integration. Therefore, it is possible t hat a female was most groomed by her bonding partners in response to socioec ological stress, providing shor t-term stress-relief to the groomed female (although uncaptured via measur es of fecal cortisol) and longer-term stress relief to the groomer (i .e., bonding partner) via the ex pectation of the benefits provided by the strengthened relationship. The Influence of Reproductive Condition on Cortisol Levels As expected, female cortisol levels vari ed with reproductive condition. In support of the consensus that late pregnancy results in the increased production of GC s (Ziegler et al. 1995; Lockwood et al. 1996; Smith and French 1997; Cavigelli 1999; Weingrill et al. 2004; Bales et al. 2005; Gesquirer et al. 2008), the pregnant females in my study had the highest baseline cortisol levels. J anes high baseline cortisol value was most likely a sole repercussion of pregnancy, as t he final month of her pregnancy (and of her disappearance) was an outlier, the primar y cause of her elevated baseline value (Appendix E). Similarly, Tinkerbells pr egnancy factored heavily into her elevated baseline cortisol value, but was not the only attributing factor. The GC levels of pregnant females are expected to rise only during late pregnancy (Ziegler et al. 1995; Lockwood et al. 1996; Smith and French 1997; Cavigelli 1999; Bales et al. 2005). Tinkerbells cortisol values, however, in creased from mid-pregnancy, indicating the
194 presence of outside stressors. Further, pr ior to her pregnancy, she demonstrated two months of elevated cort isol levels (2 to 3 times baselin e value) (Appendix E). I attribute Tinkerbells high cortisol concentrations to a combination of factors, namely her pregnant condition, low status and non-existent adul t social network. In fact, in a metaanalysis of both New and Old Wo rld primates, the lack of social support was found to be a primary predictor of high cortisol levels of subordinate individuals (Abbott et al. 2003). Tinkerbell was quantitatively most stressed from September to January, the time of year that, in part, corresponds to not only the transition season (characterized by a low and unpredictable food supply), the most stre ssful season for the females in general (Chapter 3), but also to her late pregnancy (Appendix E). Thus, during an ecologically stressful time span for a female of low rank (and thus with minimal access to high quality food resources) and with minimal soci al support, Tinkerbell had the additional physiological burden of energetic ally sustaining pregnancy. As opposed to studies that found cycling fe males to have lower GC concentrations than pregnant or lactating females (Ziegler et al. 1995; Weingrill et al. 2004), the current study found estrus cycling to be the median condi tion in terms of cort isol concentration. As expected, however, females in estrus had (non-significantly) higher cortisol levels than did cycling females not in estr us. Given that lactation-induced hyporesponsiveness of the HPA axis acts as a protective physiological mechanism for both mother and infant (Slattery and Neumann 2008), it was not surprising that lactating females demonstrated the lowest cortisol values. With female stress response suppressed by some mechanistic action of la ctation, however, I expected there to be a reduced effect as the infants aged (to year one) and nursing frequency decreased. This
195 was not the case, however. After controlling fo r season, maternal cortisol levels did not progressively increase as lactation demand decreased. I predict that the physical presence of the infants, as well as thei r continued (albeit decreased) suckling, was sufficient to maintain maternal HPA ax is hyporesponsiveness, as high levels of circulating GCs in the mothers milk would still be maladapt ive to infant development. Alternatively, but not necessar ily mutually exclusive from t he previous prediction, it may be that decreased lactation dem and does result in a decreased suppressive effect of maternal stress but that a new balance is achieved whereby cortisol levels are not significantly affected. As infants become more independent and nurse less, for example, the energetic and physiological demands of infant caretaking decrease, thereby reducing maternal stressors. Ther efore, as infants age, lactation-induced suppression of stress responsivity may decr ease, but so do potential stressors, thereby maintaining maternal GC conc entrations at a low level. The Influence of Reproductive C ondition on Affiliative Behavi or In terms of proximity and the rates at which females received grooming, data support the hypothesis that fema le reproductive condition affects affiliative behavior. Lactating females (i.e., those with dependent infants) were in close proximity to and received the most grooming from other females, reinforcing the strong attractivity effect of young infants (Chapter 4; Henzi and Barrett 2002); cycling females were the focus of male grooming effort. Interestingly, when in estrus, cycling females were observed to be groomed only by the alpha male; when these females became pregnant, however, the alpha male ceased all grooming interest. This finding provides empirical evidence supporting the conclusions in Chapter 4 that males demonstrated a cl ear preference for the females that were most likely to concei ve and possibly use grooming to bias female
196 mate choice in their favor, and that the dominant male cannot rely solely on his status to acquire matings. There was no evidence link ing the immediate exchange of grooming for sex, and so this dynamic might involve a longer-term tradeoff. The turnover rate of this trade-off, if one exists at all, would be of interest for future study. Although there was no overall effect of reproductive state on the rates at which females groomed adult troop-mates, intriguing within-condition patterns emerged. For example, that estrus females dedicated nearly all of their grooming effort to the alpha male clearly depicts female mate choice (Janson 1984). Additionally, one of the pregnant females, Jane, demonstrated an intere sting pattern to her grooming behavior, possibly as a means to expand her social network in preparation for her newborns arrival. The rate at which Jane groomed fe males strongly correlated with her length of gestation, and in her final month of pregnan cy, Janes intensified grooming effort uncharacteristically focused on the alpha male and dominant female. I interpret this as a means to strengthen ties with the individua ls that would soon become most involved and/or important in her and her babys life. Not only does this indicate an ability to socially strategize, but also the ability to strategically plan ahead. Tinkerbell, the other pregnant female, did not demonstrate such effo rt, possibly because of her virtual lack of social interaction with adult troop-mates. A second possibi lity, however, evokes the importance of experience with pregnancy and post-natal care in being able to understand what to plan ahead for. As a mu ltiparous female, Jane likely was able to call upon previous experience to comprehend her current and future situation, whereas Tinkerbell, a primiparous female, was unable to do so. While still conjecture at this
197 point, the possibility is supported by thes e initial data patterns and warrants further investigation. Just as Janes social investment was affected by gestation length, the social behavior of lactating females was associated with the length of their reproductive condition. Mothers steadily decreased groo ming behavior towards females as their infants aged, suggesting a social and/or physi ological significance to their grooming investment. Likely involved in this dynamic is the use of grooming as a commodity to be traded and/or the stress-reducing properties of the receipt of grooming, whereby the grooming of curious female neighbors by the infants mothers promoted an immediately more relaxed/less potentially agonistic env ironment surrounding the infant. As the infants aged, becoming more independent and le ss vulnerable (and less attractive to female conspecifics), the mothers were able to decrease their grooming investment in females. Additionally, that female-female proximity sco res decreased with infant age, while female-male proximity scores did not, further exemplifies that strong female-female relationships are temporar y and situation-dependent, while female-male relationships are more benefic ial and thus more stable. Highlighting the potent ial benefits of female-male rela tionships, lactating females most groomed males when their infants were less than 3 months old and as the infants approached one year of age. I predict that the former prim arily secures benefits for the newborns (e.g., socialization of the infants with the dominant male and increased tolerance and protection from predators and troop-mates) (Chapter 4), while the latter is an assertion of the females relationships wi th males as they resume hormonal cycling. Alternatively, most infants were 9 to 12 m onths of age during the transition season, and
198 thus seasonality may be a confounding factor. As already discussed, the transition season was a stressful time characterized by an unpredictable food supply (Chapter 3), and lactating females may have increased groom ing investment in males during this time as a means to strengthen their relati onships and garner the benefit of increased protection from predators and tolerance ar ound food resources (especially important when still producing energetic ally expensive milk) and/or, as previously discussed, physiologically mediate their incurred st ress via grooming and t he expectation of returned benefits. Conclusion Do females use affiliative behavio r to mediate stress? Overall, female GC concentrations were not associated with rate s of affiliative behavior or with monthly shifts in social relationships; the lack of st atistical association, however, may be due to the substantial interand intra-in dividual variation in cortisol levels. Emergent patterns provide indication of the social mediation of female stress and i dentify more focused research questions for future study. In primates, so cial bonds are not hormonally determined, but are physiologically rewarding (Curley and Keverne 2005; Dunbar 2009). If there is a causal basis to the inverse a ssociation between female cortisol levels and the grooming of adult males, then this underscores that female C. apella in RV, despite being female-philopatric, are more bonded to males than to females. That females received more grooming from strongly-bonded troop-mates during periods of increased GC concentration supports the notion that smaller, more concentrated grooming networks have a greater impac t on stress reduction than do larger, more diverse ones (Wittig et al. 2008). Further evidence supporting the possible dynamic between affiliative relationships and stress management is the finding that the most subordinate
199 female demonstrated an obvious lack of adul t social support and had the highest overall GC concentration. Female cortisol levels did vary by r eproductive condition, and female reproductive condition was found to affect female affiliative behavior. Elucidating a direct relationship between reproductive stress and social beha vior, however, was beyond the scope of my study and requires a focused investigation in itself. The dynamic between stress hormones and pregnancy is particularly complic ated, as the expected (and necessary) peripartum increase in GCs does not prec lude pregnant females from experiencing actual stress, nor does it negate the internal stressor of being pregnant, especially when confounded by other ex acerbating factors such as lo w rank, lack of social support, and/or ecological stress. My study, however, provided an initial foray into untangling such variables, and an intriguing possibility em erged: An experienced (i.e., multiparous) pregnant female may have understood her current and future reproductive situation and was able to plan ahead and socially strategi ze, focusing her grooming effort on the individuals that would become more important and/or involved in her and her infants life, possibly as a means to circumvent so me of the stress involved in pregnancy and infant-caretaking. Further, despite hypor esponsiveness of the maternal HPA axis, lactating females with young infants experienced increased co rtisol levels, likely in response to unsolicited attention (namely fr om female troop-mates) that was beyond the mothers control. The increased rate of grooming that mothers devoted to female neighbors during this time can be interpreted as an immediate means to promote a relaxed social setting, essentially protecting the infants from risk of escalated agonism.
200 Table 5-1. Summary of the dominance r ank, approximate age, bi rthing history and reproductive classification (for pur poses of my study) of the adult C. apella focal females. Female Rank Est. age (years) Birthing history (most recent birth, as of Dec 2006) Reproductive Classification*(Jan-Dec 2006) Notes of interest Gina 1 > 10 Multiparous (Dec. 24, 2005) Lactating Banana 2 6 Primiparous (Jan. 21, 2006) Lactating Jane 3 8 Multiparous (Feb. 22, 2005), pregnant Cycling, Pregnant Disappeared in October Suspected obstetricrelated death (near full term) Little Horns 4 > 15 Multiparous (Dec. 6, 2004) Cycling Carol 5 > 10 Multiparous (Nov. 22, 2005) Lactating Kate 6 > 10 Multiparous (Feb. 5, 2006) Lactating Tinkerbell 7 6 Nulliparous, pregnant Cycling, Pregnant Gave birth to 1st infant after conclusion of study Females typically spanned at least two diff erent reproductive conditions during the study period; their dominant condition(s) (whi ch they most represent in the data set) is/are listed. Table 5-2. Females mean monthly values ( SE) of cortisol (F), total proximity score (TPS), total female proximity score ( TFPS) and male proximity score (MPS). Overall, there was no troop-wi de association between femalesX proximity scores and cortisol levels (TPS: rs = -0.43, P = 0.34; TFPS: rs = -0.64, P = 0.12; MPS: rs = -0.36, P = 0.43). Female F TPS TFPS MPS Gina 48.29 15.36 214.9 28.83 52.7 15.47 41.1 11.54 Banana 76.05 16.06 215.6 45.85 38.4 13.10 39.3 11.82 Jane 90.21 22.89 69.1 14.91 5.6 2.37 21.5 10.72 Little Horns 67.70 12.63 63.2 12.43 1.5 0.79 14.3 3.69 Carol 69.63 19.26 105.7 15.86 10.7 4.91 12.4 3.64 Kate 54.83 21.83 79.5 29.25 11.6 5.83 12.2 5.90 Tinkerbell 100.65 27.31 47.3 10. 88 0.7 0.68 9.5 4.96 Troop X 72.15 7.33 113.61 27.08 17.31 7.62 21.47 5.04
201 Table 5-3. Range of values for each females X monthly cortisol levels and total monthly proximity to females (TFPS) and to males (MPS). There were no significant associations between cortisol and proximity for any focal female. Female Cortisol (ng/g) TFPS TFPS association MPS MPS association Gina 10.9 214.1 2.4 155.5 rs=0.20, P=0.61 4.3 103.0 rs=0.38, P=0.31 Banana 16.4 188.3 0.0 133.0 rs=0.07, P=0.85 3.6 123.7 rs=0.30, P=0.40 Jane 26.4 321.6 0.0 21.5 rs=0.08, P=0.83 0.0 95.6 rs=0.58, P=0.10 Little Horns 18.5 127.8 0.0 7.8 rs=0.42, P=0.22 0.0 37.9 rs=0.39, P=0.26 Carol 19.5 238.3 0.6 31.7 rs=0.24, P=0.51 0.0 26.6 rs=0.14, P=0.70 Kate 8.7 274.1 0.0 58.3 rs=0.06, P=0.88 0.0 54.6 rs=0.13, P=0.73 Tinkerbell 16.0 245.2 0.0 6.8 rs=-0.06, P=0.87 0.0 47.8 rs=-0.31, P=0.38 Table 5-4. Association between a females frequency of occupying a central position in the troop and her social proximity to adult females (TFPS) and adult males (MPS). Most females showed a positive relationship between centrality and proximity scores, although only the two most subordinate females (Kate and Tinkerbell) demonstrated significant relationships. Female TFPS MPS Gina rs = 0.58, P = 0.08 rs= 0.44, P = 0.21 Banana rs = 0.62, P = 0.06 rs= 0.16, P = 0.66 Jane rs = -0.12, P = 0.76 rs= 0.41, P = 0.27 Little Horns rs = -0.22, P = 0.55 rs=0.12, P = 0.73 Carol rs = 0.11, P = 0.77 rs= -0.33, P = 0.34 Kate rs = 0.97, P < 0.0001 ** rs= 0.71, P = 0.02 ** Tinkerbell rs = 0.59, P = 0.07 rs= 0.77, P = 0.01 ** Notes: **significant relationship ( P < 0.05); relationship approaching significance (0.05 < P 0.08)
202 Table 5-5. Relationship between months with and without prox imity strong bonds andX monthly cortisol (F) levels (ng/ g). No female had a significant association between the presence of pr oximity strong bonds and cortisol levels, nor was there a consistent pa ttern among the females in terms of whether their X cortisol levels were higher in months with versus without strong proximity bonds. Female nwith X F SE nwithout X F SE P -value F levels higher Gina 9 61.6 22.9 0 ---Banana 9 77.5 18.6 1 93.4 -0.79 without bonds Jane 5 103.3 34.4 4 105.3 75.1 0.98 same Little Horns 4 58.4 23.9 6 58.0 13.3 0.99 same Carol 5 98.6 40.3 5 60.3 23.3 0.44 with bonds Kate 5 87.2 49.1 4 44.7 27.9 0.51 with bonds Tinkerbell 3 50.9 31. 6 7 114.7 38.9 0.35 without bonds Table 5-6. Relationship between months with and without grooming (given) strong bonds andX monthly cortisol (F) levels ( ng/g). One female, Banana, had significantly higher cortisol levels in months when she did not express any strong grooming bonds, while two fema les, Gina and Carol, had nearly significantly higher cortisol levels in months when they did express at least one strong dyadic grooming bon d. Overall, there was no consistent pattern among female groomers in terms of whether their X cortisol levels were higher in months with versus without strong gr ooming bonds. Female nwith X F SE nwithout X F SE P -value F levels higher Gina 7 74.8 27.7 2 15.5 4.6 0.08 with bonds* Banana 8 59.7 11.9 2 156.8 31.5 0.04 without bonds ** Jane 8 91.3 36.4 1 207.2 -0.25 without bonds Little Horns 6 59.6 12.7 4 55.9 24.6 0.52 same Carol 8 94.1 26.2 2 20.7 1.2 0.07 with bonds* Kate 5 33.3 16.6 4 111.9 59.4 0.14 without bonds Tinkerbell 0 -10 95.6 29.5 --Notes: **significant relationship ( P < 0.05); relationship approaching significance (0.05 < P 0.08)
203 Table 5-7. Relationship between months with and without grooming (received) strong bonds andX monthly cortisol (F) levels (ng/g) Despite a lack of significant associations, a descriptive trend emerged: All females (with the exception of the most subordinate female, Tinkerbell) had higher cortisol levels in the months characterized by the presence of strong grooming bonds (grooming received). Female nwith X F SE nwithout X F SE P -value F levels higher Gina 5 77.0 38.0 4 42.4 22.9 0.62 with bonds Banana 5 105.2 26.25 53.0 15.0 0.08 with bonds Jane 3 167.7 82.96 74. 5 29.9 0.12 with bonds Little Horns 7 66.9 14.9 3 37.8 13.6 0.21 with bonds Carol 4 114.1 50.06 56.3 17.2 0.40 with bonds Kate 4 101.8 60.65 41.5 21.8 0.81 with bonds Tinkerbell 2 19.3 0.8 8 114. 6 33.7 0.30 without bonds Note: relationship approac hing significance (0.05 < P 0.08) Table 5-8. Effect of infant age on mothers ( n = 4) monthly proximity to adult females (TFPS) ( n = 6) and adult males (MPS) ( n = 3). As the infants aged and became more independent, there was an associated decrease in the mothers proximity to other females ( rs = -0.44, P = 0.005). No such association, however, resulted between infant age and mothers proximity to males ( rs = 0.06, P = 0.74). There were no statis tical differences between TFPS and MPS at any of the increment al infant age categories ( P = 0.09 to 0.59), but it is interesting to note that for mo thers with dependent infants, TFPS remained higher than MPS for the first 6 months of their infants lives, after which MPS was higher than TFPS. Infant age (months) n TFPS SE MPS SE 0.0-3.0 11 47.15 13.50 27.59 9.15 3.1-6.0 11 35.61 13.26 29.92 11.99 6.1-9.0 9 17.24 8.62 23.41 8.81 9.1-12.0 8 8.30 3.78 25.65 9.67
204 Table 5-9. Monthly association between gesta tion length, proximity to females (TFPS) and proximity to males (MPS). Pregnant females ( n = 2) maintained a higher (non-significant) MPS t han TFPS during each month of pregnancy. There was no relationship between gestation length and proximity to females ( rs = -0.09, P = 0.81) or males ( rs = 0.24, P = 0.51). Month of pregnancy (approximated) n TFPS SE MPS SE P value 1 2 5.5 5.5 17.9 17.0 0.48 2 1 0.0 3.6 n.a. -3 1 0.0 6.8 n.a. -4 2 0.0 1.6 0.75 0.29 5 2 10.8 10.8 24.0 20.6 0.41 6 2 4.3 4.3 5.6 0.7 0.77 Figure 5-1. Troop-level comparison of female reproductive state and X baseline cortisol level SE (ng/g). Overall, pregnant females had the highest cortisol levels (X =75.29 17.01), while lactating females had the lowest cortisol levels (X = 43.26 4.90). Posthoc analysis revealed that lactating females had significantly lower cortisol levels than anestrous ( P = 0.027), estrus ( P = 0.018) and pregnant (P = 0.033) females. 0 10 20 30 40 50 60 70 80 90 100anestrous estrus pregnant lactating** anestrous (n = 43) estrus ( n =28 ) pregnant (n = 35) lactating (n = 131) Cortisol (ng/g)
205 Figure 5-2. Correlation between Tinkerbells cortisol levels (ng/g) and month of pregnancy. Her cortisol levels in creased steadily throughout her pregnancy and were strongly and positively correlated with gestation length ( rs = 0.94, P = 0.005). 0 50 100 150 200 250 300 0.51.52.53.54.55.5 Month of gestation (estimated) Cortisol (ng/g)
206 Figure 5-3. Comparison of Janes mean cortis ol levels SE (ng/g) during months of her pregnancy. In October, her final month of pregnancy (also the month of her presumed death), Janes mean cortisol level was significantly higher than in all other months in which she was pregnant (and baseline cortisol data available). Figure 5-4. Cortisol levels of lactating females ( n = 4) at incremental stages of infant age. Cortisol levels were highest for la ctating females when their infants were 3 months of age and lowest when their in fants were 6 to 9 months of age. The cortisol levels of lactating female s were significantly higher when their infants were 9 to12 months old as compared to 3 to 6 months ( P = 0.050) and 6 to 9 months (P < 0.0001), and when the infants were 0 to 3 months as compared to 6 to 9 months ( P = 0.028). 0 50 100 150 200 250 300 350 JunAugSeptOct 0 10 20 30 40 50 60 70 80 90 100 0-3 months3-6 months6-9 months9-12 monthsCortisol (ng/g)Infant ageCortisol (ng/g) **
207 Figure 5-5. Rates ( SE) at which cyc ling (black bars), pregnant (grey bars), and lactating (white bars) females gr oomed and received grooming from adult female and male troop-mates. R eproductive condition did not have a significant effect on the rates at which females groomed females ( P = 0.293) or groomed males ( P = 0.196). There was, however a significant effect on the rates at which females of different reproductive condition were groomed by females ( P = 0.010), and a nearly significant effect on the rates at which such females were groomed by males ( P = 0.056). In each reproductive condition, females groomed males ( n = 3) at a higher ( non-significant) rate than they groomed females (n = 7). In turn, lactat ing females received a significantly higher rate of gr ooming from females than males ( P = 0.012), while cycling females received a higher (non-significant) rate of grooming from males than females. 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05 FemalesMalesFemalesMales Groom Receive groom fromGrooming rate (min/h)
208 Figure 5-6. Effect of infant age (3-month increments) on the ra te ( SE) at which mothers (i.e., lactating females) groomed and received grooming from adult females and males. Although there was no significant effect of infant age on the rate at which mothers groomed females ( P = 0.249), grooming decreased steadily as infants became more dependent, and the two variables were moderately correlated ( rs = -0.37, P = 0.021). Mothers mo st groomed males when their infants were 0 to 3 and 9 to 12 months old as compared to the other infant age increments, and this difference approached significance ( P = 0.061). Although infant age did not signifi cantly affect the rate at which mothers received grooming from females ( P = 0.107) or males ( P = 0.420), mothers received most groom ing from females when their infants were 0 to 3 months old. 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0-33-66-99-120-33-66-99-120-33-66-99-120-33-66-99-12 Groom femaleGroom maleGroom by femaleGroom by maleGrooming rate (min/h) Infant age (months)
209 CHAPTER 6 CONCLUSION The central premise of my research address ed the intrinsic effects of affiliatio n and the social, reproductive and ecological circumst ances that may prompt shifts in female relationships. Laboratory studies indicate a reciprocal relationship between stress (as measured by glucocorticoid levels) and social support, but typically do so in an environment that controls for the natural range of social and ecological variables that characterize an animals evolutionary hi story. Female brown capuchins at Raleighvallen (RV), Suriname are faced with t he complex challenges of residing in an unpredictable environment, one in which the cost/benefit tradeoffs of sociality can quickly shift. The purpose of my study was to determine how females adapt to such challenges, both behaviorally and physiologically. In order to evaluate female response to stress, I first ha d to identify the socioecological variables that act as prim ary stressors to females. For group-living species, it is generally accept ed that the main cost of soci ality is increased within-group competition, while the main benefit is improved predator protection (Isbell 1994; Sterck et al. 1997). The cost/benefit regime is mu ch more complicated, however, and adjusts in response to changes in local and temporal circumstances. The dynamics of group living are further exacerbated by group size, as the costs and benefits of sociality are differently balanced for large versus small groups. I studied two capuchin troops of disparat e size (representing opposite ends on the group size continuum scale) and found that fe males in the larger group (Troop A) incurred more of the costs of sociality (i .e., increased agonism and greater ecological demands) but did not receive the benefits of a large social group. In fact, Troop A
210 females received less grooming than did females in the small group (Troop B), despite a three-fold difference in the possible number of social partners. Further, the large social setting did not appear to alleviate perceived predation risk, but did affect the source (aerial versus terrestrial) of their greatest perceived risk (Chapter 2). Fecal cortisol levels supported these findings; overall, fema les in Troop A had higher cortisol levels than did the females in Troop B, mostly in response to aerial predators and an unpredictable food supply (Chapter 3). With the intense ecological variati on, both within and between years, and the strong presence of both aerial and terrestrial predators that characterizes the RV forest, the expectation of shifts in the cost/benefit r egime for a group of any given size must be enough to maintain disparate group size. De spite the prolonged consequences felt by Troop A females during the study period, the risks of emigrating must have been greater and/or they expected the scale to soon tip in their favor. In such a dynamic ecological landscape, the ability to adapt is most adaptiv e. As any natural environment of wild living social mammals is more dynamic than st able, the theoretical notion of an optimal group size (Silk 2007a) is rather unrealisti c and may only apply to captive populations. After establishing the foundation for inte rgroup variation in the socioecological lives of female brown capuchins in RV, I sh ifted my focus to study ing more fine-scaled within-group social behavior (Chapter 4) and the potentia l physiological benefits of affiliation (Chapter 5). Because most group-living nonhuman primates are female-philopatric, it is often assumed t hat relationships among females are strong and adaptive. In fact, the concept of female bonds is at the core of primate social evolution theory (Wrangham 1980). Although capuchins are supposedly a female-bonded genus,
211 with females demonstrating stronger relationships with females than with males (Perry 1996; Di Bitetti 1997), the results of my study indicated that, overall, the females in RV are not female-bonded but are better characterized as cross-bonded with adult males. The term cross-bonded is commonly used to describe baboon females (Byrne et al. 1990; Barton et al. 1992; Swedell 2002) but has not been previously used to describe any capuchin group or population. Despite being linked to their female troop-mates through kinship and a long histor y of familiarity, proximit y and grooming relationships between females were transient and most predi ctably occurred in response to an infant. Female relationships with males were mo re stable and appeared to be related to the acquisition of increased tolerance and protection-commodities that are especially important in the high risk, lo w energy environment of RV. Female-female and femalemale relationships are strategic responses to ecological and reproductive demands and thus cannot be generalized at the genus, or even species level (Chapter 4). Regardless of whether a females primary social partners are ma le or female, the proximate mechanism underlying the relationshi p remains: Primate social bonds are said to be physiologically rewarding (D unbar 2009) based, in part, on the potentially stress-reducing properties of gr ooming (Boccia et al. 1989; Ke verne et al. 1989; Gust et al. 1993; Aureli et al. 1999; Shutt et al. 2007). Thus, the elevated cortisol levels of Troop A females (Chapter 3) likely resulted fr om an interplay of the direct and indirect consequences of larger group size. Since their intensified foraging demands reduced the time available to rest and groom, t he females in Troop A had less opportunity to benefit from grooming than di d the Troop B females.
212 As prolonged stress deleteriously affects the health and fitness of females and their developing offspring (Sachser et al 1998; Bardi et al. 2005; Sapolsky 2005; Shively et al. 2005; Wadhwa 2005), females should take an active role in managing their stress through the strat egic use of grooming. Alt hough the physiological benefits of the receipt of grooming occur immediatel y upon tactical contact and thus are difficult to capture using measures of longer-term fecal cortisol (Shutt et al. 2007), my study presents anecdotal evidence and empirical patterns that implicate an association between affiliation and stress mediation. For example, females with young infants ( 3 months old) may have int ensified their grooming of curi ous conspecifics in order to promote an immediately more relaxed social environment surrounding their vulnerable infants. Females also tended to be gr oomed more intensely by specific adult troop-mates during periods of increased stress, a system of social support similar to that found in humans (Kirschbaum et al. 1995; Taylor et al. 2000; Heinrichs et al. 2003; Rosal et al. 2004) and some captive mammals (Mendoza et al. 1991; Carter et al. 1995; Sachser et al. 1998; von Holtz 1998; Ru kstalis and French 2005; Hennessey et al. 2006). Further, the results of my study support the novel finding t hat the giving of grooming is associated with a decrease in cortis ol levels (Shutt et al 2007). Shutt et al. (2007), however, found such an association within female-female grooming dyads, whereas in my study, it was the female groomi ng of males that was specifically linked to lower cortisol concentrations. Not only does this finding support the notion that female brown capuchins in RV are more strongl y bonded to males than to females, but it identifies a possible motivating factor for the maintenance of social bonds (Chapter 5).
213 The present research (1) ex pands the breadth of known in traspecific variation in behavioral responses to ecological diversit y of a social mammal; (2) provides an operationalized approach to the study of social bonds and ident ifies issues that limit the comparative application of current published studies; and (3) contributes the first investigation into the behavioral endocrinology of wild female brown capuchin monkeys, strengthening the known eviden ce linking social bonds to fitness consequences. The incidence of stress-related disease is incr easing in both human (Tamashiro et al. 2005) and captive animal populations By providing an improv ed understanding of natural stressors and the adaptive implications of social evolution, my study has direct ramifications for the health and wellbeing of humans and nonhumans alike.
214 APPENDIX A CHAPTER 2 RESULTS SUMMARY
215 Variable Troop A Troop B Statistical result Association between group size and variable Troop size Density (# indviduals/m) (based on X troop dispersion) # of # of contact hours 27 0.4 7 2284 9 0.2 3 495 Home range size Home range composition (%) Bamboo High forest Liana/low forest Swamp forest Habitat preference (time %) 700 m 2 13 54 28 5 Bamboo (37%) 500 m 2 10 44 43 3 Liana/low forest (49%) X daily travel rate (m/h) Most traveled season Least traveled season 171 Transition Wet 159 Transition Wet A > B ( P = 0.037)* (+) X canopy cover (0-3 incremental scale; 0=fully exposed; 3=full cover) 2.41 2.34 A = B none
216 Appendix A. Continued. Variable Troop A Troop B Statistical result Association between group size and variable X height in canopy Lower limit (m) Upper limit (m) 4.95 10.97 6.50 12.23 B > A ( P < 0.0001) B > A ( P < 0.0001) (-) X monthly rate of M. maripa feeding (min/h) Troop rate Individual rate 4.36 0.37 5.70 1.26 A = B B > A ( P = 0.020) (-) X monthly rate of AAs (alarms/h) Troop rate Individual rate 0.62 0.03 0.66 0.09 A = B B > A ( P < 0.0001) (-) X monthly rate of TPAs (alarms/h) Troop rate Individual rate 0.17 0.008 0.07 0.010 A > B ( P = 0.010) A = B (+) X monthly individual rate of agonism (bouts/h) as aggressor as victim 0.009 0.009 0.004 0.002 A > B ( P = 0.004) A > B ( P = 0.001) (+) agonism (% of intragroup agonism hierarchy 11.8 Present, weakly linear 0.0% Absent (+) X monthly individual rate of grooming (min/h) as groomer as recipient 0.14 0.06 0.19 0.13 A = B A = B (-)
217 APPENDIX B ADDITIONAL RESULTS PERTA INING TO TABL ES 4-5 AND 4-6: DYADIC PATTERNS OF FEMALE AFFILIATION (TROOP A) Proximity Ginas proximity scores significantly differed by adult dyadic partner ( F = 8.489, P < 0.0001, df = 8), and this difference was due to her proximity to Boris, the alpha male (mean S E = 27.25 8.27), and Banana, the 2nd ranking female (mean SE = 23.37 4.63). Ginas scores were signi ficantly higher with Boris ( P = 0.000 to 0.001) and with Banana ( P = 0.000 to 0.013) than wit h all other adults (mean SE = 0.15 to 5.31). There was no significant difference betw een Ginas proximity scores with Boris and Banana ( P = 0.997). Bananas proximity scores significantly differed by adult dyadic partner ( H = 46.50, P < 0.0001, df = 8), and this difference was due to her proximity to Gina (mean SE = 23.33 4.65) and Boris (mean SE = 15. 17 3.70). Bananas scores were significantly higher with Gina (P = 0.000 to 0.003) and with Boris (P = 0.000 to 0.015) than with all other adults (m ean SE = 0.02 to 5.44). There was no significant difference between Bananas proximity scores with Gina and Boris ( Z = -1.209, P = 0.247). Janes proximity scores significantly differed by adult dyadic partner ( H = 22.364, P = 0.004, df = 8), and this difference was due to her proximity with the alpha male. Janes proximity to Boris (mean SE = 12.67 6.39) was significantly higher than with most adult social partners: Ti nkerbell (mean SE = 0.0 0.0, P = 0.004), Little Horns (mean SE = 0.29 0.19, P = 0.008), Kate (mean SE = 0.47 0.27, P = 0.008), Andycap (mean SE = 0.48 0.36, P = 0.011), Carol (mean SE = 0.49 0.28,
218 P = 0.011), and Banana (mean SE = 0.77 0.27, P = 0.024), except for Gina (mean SE = 4.88 2.81, P = 0.190) and Darwin (mean SE = 9.17 5.81, P = 0.222). Little Horns proximity scores signific antly differed by adult dyadic partner ( H = 31.508, P < 0.0001, df = 8), and this difference was due to her proximity to the males. Little Horns dyadic proximity scores were significantly higher ( P < 0.0001) with adult males (mean SE = 3.84 0.98) t han with adult females (mean SE = 0.32 0.13). There was no significant differenc e in her dyadic proximity scores among the males (means = 2.30 to 5.06, P = 0.839) or among females (means = 0.07 to 0.68, P = 0.238). Carols proximity scores did not signifi cantly differ by adult dyadic partner (means = 0.14 to 4.98) ( H = 12.033, P = 0.150, df = 8); the individuals with whom she scored the highest average proximity scores were both males: Andycap (mean SE = 4.98 2.88) and Darwin (mean SE = 4.85 3.40). Kates proximity scores significantly differed by adult dyadic partner ( H = 19.445, P = 0.013, df = 8), but there was no discernable patte rn to the differences. Although her highest average proximity score was with Bo ris (mean SE = 8.39 3.99), only the mean rank of Kates proximity scores wit h Carol (mean SE = 4.61 2.31) and with Banana (mean SE = 5.44 3.18) were si gnificantly higher than with Tinkerbell (mean SE = 0.07 0.05, Pboth = 0.005), Little Horns (mean SE = 0.18 0.13, Pboth = 0.015), and Andycap (mean SE = 0.38 0.26, PCarol = 0.029, PBanana = 0.043). Tinkerbells proximity scores signific antly differed by adult dyadic partner ( H = 18.367, P = 0.019, df = 8), and this difference was due to her proximity to Andycap (mean SE = 1.43 0.69). Although Tinker bells highest average dyadic score was
219 with Boris (mean SE = 11.0 5.47), monthly measures of her proximity to the alpha male varied drastically (SD = 17.31) and thus t he mean value is an unreliable predictor. Only the mean rank of Tinkerbells proximity score with Andycap was significantly higher than with Jane (mean SE = 0.00 0.0, P = 0.028) and with Carol (mean SE = 0.05 0.05, P = 0.043). Grooming Gina was observed to groom all of the adul t females exc ept Tinkerbell, although 86.5% of the bouts were devoted to Banana. Of the three adult males, Gina was observed to only groom the alpha male. Over all, Gina dedicated the most grooming to Boris (48.6% of bouts and 48.7% of total duration) and Banana received the remaining majority of her attention ( 44.4% of grooming bouts and 47.1% of total duration). Banana was observed to groom all of the adult females except Jane and Carol, although 56.9% of the bouts were devoted to Gina. Of the three adult males, Banana was observed to groom Boris almost exclusiv ely. Overall, Banana groomed Gina most frequently (37.7% of bouts), but she groomed Boris for the longest duration (50.8% of total grooming time). Jane was observed to groom all of the adul t females except Kate and Tinkerbell, with Little Horns receiving 45% of the bout s. Of the three adult males, Jane was observed to groom only Boris. Overall, Jane dedicated the most grooming to Boris (45.9% of bouts and 45.0% of total duration). Little Horns was observed to groom all of the adult females except Tinkerbell, with no single female receiving a distinct majori ty, and she groomed all three adult males. Little Horns groomed a male during 54.5% of her grooming bouts, and males received
220 58.6% of her grooming time. Overall, Li ttle Horns dedicated the most grooming to Darwin (39.4% of bouts and 49.4% of total duration). Carol was observed to groom all of the adul t females and males of the group; the groups six females received 54.5% of her grooming bouts, but the three males received 65.2% of Carols grooming time. Overall, Carol dedicated the most grooming to Boris (30.3% of bouts and 29.1% of total duration). Kate was observed to groom all of the adult females except Gina and Jane, with Banana receiving 75% of the bouts. Of the three adult males, Kate was observed to groom only Boris and Andycap. Overall, Ka te dedicated the most grooming to Boris (41.7% of bouts and 64.6% of total duration). Tinkerbell rarely participated in grooming with other adults. She was observed to groom only Banana, Carol and Darwin, and onl y for a total of 3.8 min (4 bouts) throughout the study period. In fact, Tink erbell was never observed to groom or be groomed by five of the nine adult troop members. Boris was observed to groom all of the adult females except Tinkerbell. He devoted the most grooming attention to Jane (29% of bouts and 32.6% of total duration). Andycap groomed only Little Horns and Carol, with Carol receiving 92.5% of his total grooming duration. Darwin, on the other hand, groomed all of the females except Banana, with the vast majority devoted to Little Horns (54.5% of bouts and 59.2% of total duration).
221 APPENDIX C TOTAL PROXIMITY SCORES (PS) FOR ALL FEMALE-FEMALE AND FEM ALE-MALE DYADS (TROOP A) (JANUARY-DECEMBER, 2006) Focal Dyad partner Total PS Gina Boris 272.492 Gina Banana 233.69 Banana Gina 233.27 Banana Boris 151.674 Jane Boris 114.048 Tinkerbell Boris 110.026 Kate Boris 83.929 Jane Darwin 82.56 Banana Kate 54.399 Kate Banana 54.399 Gina Darwin 53.135 Banana Darwin 51.857 Little Horns Darwin 50.604 Carol Andycap 49.843 Carol Darwin 48.461 Banana Andycap 46.414 Carol Kate 46.103 Kate Carol 46.103 Jane Gina 43.932 Little Horns Boris 41.648 Gina Kate 31.773 Kate Gina 31.773 Banana Jane 28.163 Gina Jane 25.551 Carol Boris 23.847 Gina Carol 23.409 Little Horns Andycap 23.013 Carol Gina 21.349 Kate Darwin 19.972 Gina Little Horns 18.615 Tinkerbell Andycap 14.299 Gina Andycap 14.1 Tinkerbell Banana 11.151 Banana Carol 10.105 Carol Banana 10.105 Strong bonds (top 10% of scores)
222 Appendix C. C ontinued. Focal Dyad partner Total PS Banana Tinkerbell 8.536 Jane Banana 6.893 Little Horns Gina 6.801 Little Horns Carol 5.775 Carol Little Horns 5.775 Jane Carol 4.367 Carol Jane 4.367 Jane Andycap 4.295 Jane Kate 4.273 Kate Jane 4.273 Kate Andycap 3.789 Jane Little Horns 2.615 Tinkerbell Darwin 2.564 Little Horns Jane 1.874 Little Horns Kate 1.835 Kate Little Horns 1.835 Little Horns Tinkerbell 1.7 Tinkerbell Little Horns 1.7 Gina Tinkerbell 1.545 Tinkerbell Gina 1.545 Carol Tinkerbell 1.36 Kate Tinkerbell 0.732 Tinkerbell Boris 0.732 Little Horns Banana 0.695 Tinkerbell Carol 0.504 Banana Little Horns 0.21 Jane Tinkerbell 0 Tinkerbell Jane 0 Females (in order of decreasing rank): Gina, Banana, Jane, Little Horns, Carol, Kate, and Tinkerbell; Males: Boris (alpha), Andycap, and Darwin Lowest 10% of scores
223 APPENDIX D TOTAL GROOMING DURATIONS FOR ALL FEMALE-FEMALE AND FEMALE-MALE DYADS (TROOP A) (JANUARYDECEMBER, 2006) Groomer Recipient Total Grooming duration (min) Gina Boris 86.5 Gina Banana 83.8 Banana Boris 82.5 Kate Boris 51.2 Banana Gina 49.8 Little Horns Darwin 40.6 Jane Boris 36.25 Darwin Little Horns 30.2 Carol Boris 24.2 Carol Darwin 20 Banana Kate 18.8 Boris Jane 18.2 Jane Little Horns 17.8 Kate Banana 17.5 Boris Little Horns 16.2 Andycap Carol 16 Jane Banana 12.2 Jane Gina 12 Little Horns Gina 12 Carol Andycap 10 Carol Little Horns 9.2 Carol Banana 9 Little Horns Carol 8.5 Little Horns Boris 7.1 Darwin Tinkerbell 7 Boris Banana 6.7 Boris Kate 6.1 Little Horns Kate 6 Kate Andycap 5.3 Darwin Gina 5.3 Boris Carol 5 Darwin Kate 5 Banana Tinkerbell 4.7 Carol Gina 4.5 Strong bonds (top 10% of total durations)
224 Appendix D. Continued. Groomer Recipient Total Grooming duration (min) Little Horns Banana 4 Boris Gina 3.7 Gina Carol 3.5 Banana Little Horns 3.5 Little Horns Jane 3.5 Carol Jane 3.2 Banana Darwin 3 Kate Tinkerbell 3 Tinkerbell Banana 2.5 Jane Carol 2.3 Gina Little Horns 2 Carol Kate 2 Kate Little Horns 2 Darwin Jane 2 Darwin Carol 1.5 Andycap Little Horns 1.3 Gina Jane 1 Gina Kate 1 Carol Tinkerbell 1 Tinkerbell Carol 1 Little Horns Andycap 0.5 Tinkerbell Darwin 0.3 Kate Carol 0.2 Gina Tinkerbell 0 Gina Andycap 0 Gina Darwin 0 Banana Jane 0 Banana Carol 0 Banana Andycap 0 Jane Kate 0 Jane Tinkerbell 0 Jane Andycap 0 Jane Darwin 0 Little Horns Tinkerbell 0 Kate Gina 0 Kate Jane 0 Kate Darwin 0
225 Appendix D. Continued. Groomer Recipient Total Grooming duration (min) Tinkerbell Gina 0 Tinkerbell Jane 0 Tinkerbell Little Horns 0 Tinkerbell Kate 0 Tinkerbell Boris 0 Tinkerbell Andycap 0 Boris Tinkerbell 0 Andycap Gina 0 Andycap Banana 0 Andycap Jane 0 Andycap Kate 0 Andycap Tinkerbell 0 Darwin Banana 0 Females (in order of decreasing rank): Gina, Banana, Jane, Little Horns, Carol, Kate, and Tinkerbell; Males: Boris (alpha), Andycap, and Darwin
226 APPENDIX E GRAPHICAL REPRESENTATION OF EACH FEMALES BASELINE CORTISOL VALUE AND MONTHLY VARIAT ION ABOUT THAT MEAN (JANUARY-NOVEMBER, 2006) 0 50 100 150 200 250 JanFebMarArpJunAugSeptOctNov 0 20 40 60 80 100 120 140 160 180 200 JanFebMarAprMayJunAugSeptOctNovCortisol (ng/g) Gina Cortisol (ng/g) Banana
227 Appendix E. Continued. 0 50 100 150 200 250 300 350 FebMarAprMayJunAugSeptOct 0 20 40 60 80 100 120 140 JanFebAprMayJunAugSeptOctNovCortisol (ng/g) Jane Cortisol (ng/g) Little Horns
228 Appendix E. Continued. 0 50 100 150 200 250 300 JanFebMarAprMayJunJulAugSeptOctNovDec 0 50 100 150 200 250 300 FebMarAprMayJunJulAugSeptOctNovCortisol (ng/g) Carol Cortisol (ng/g) Kate
229 Appendix E. Continued. 0 50 100 150 200 250 300 JanFebMarAprMayJunAugSeptOctNovCortisol (ng/g) Tinkerbell
230 LIST OF REFERENCES Abbott DH, Keverne EB, Bercovitch FB Shiv ely CA, Mendoza SP, Saltzman W, Snowdon CT, Ziegler TE, Banjevic M, Garland T, Sapolsky RM. 2003. Are subordinates always stressed? A comparat ive analysis of rank differences in cortisol levels among primates. Horm Behav. 43:67-82. Ahumada JA. 1992. Grooming behavior of spider monkeys ( Ateles geoffroyi ) on Barro Colarado Island, Panama. Int J Primatol. 13: 33-49. Altemus M, Deuster PA, Galliven E, Cart er CS, Gold PW. 1995. Suppression of hypothalamic-pituitary-adrenal axis respons es to stress in lactating women. J Clin Endocrinol Metab. 80:2954-2959. Altemus M, Redwine LS, Leong YM, Frye CA, Porges SW, Carter CS. 2001. Responses to laboratory psychosocial stress in postpartum women. Psychosom Med. 63:814-821. Altmann J. 1974. Observational study of behaviour: Sampling methods. Behaviour. 49:227-267. Altmann J, Lynch JW, Nguyen N, Alberts SC, Gesquiere LR. 2004. Life-history correlates of steroid concentrations in wild peripartum baboons. Am J Primatol. 64:95-106. Aureli F, Preston SD, de Waal FB M. 1999. Heart rate responses to social interactions in free-moving rhesus macaques ( Macaca mulatta ): A pilot study. J Comp Psychol. 113:59-65. Bahr NI, Palme R, Mohle U, Hodges JK, Heistermann M. 2000. Comparative aspects of the metabolism and excretion of cortisol in three individual nonhuman primates. Gen Comp Endocrinol. 117:427-438. Bales KL, French JA, Hostetler CM, Dietz JM. 2005. Social and reproductive factors affecting cortisol levels in wild female golden lion tamarins ( Leontopithecus rosalia ). Am J Primatol. 67:25-35. Bardi M, Bode AE, Ramirez SM, Brent LY. 2005. Maternal care and development of stress responses in baboons. Am J Primatol. 66:263-278. Barrett L, Gaynor D, Henzi SP. 2002. A dy namic interaction between aggression and grooming reciprocity among female chacma baboons. Anim Behav. 63:10471053. Barrett L, Henzi SP. 2002. Constraints on relationship formation among female primates. Behaviour. 139:263-289.
231 Barton RA, Whiten A. 1993. Feeding com petition among female olive baboons, Papio anubis Anim Behav. 46:777-789. Barton RA, Whiten A, Strum SC, Byrne RW, Simpson AJ. 1992. Feeding competition among female olive baboons, Papio anubis Anim Behav. 43:831-844. Beauchamp G, Ruxton GD. 2007. False al arms and the evolution of antipredator vigilance. Anim Behav. 74:1199-1206. Beehner JC, Bergman TJ, Cheney DL, Seyfar th RM, Whitten PL. 2005. The effect of new alpha males on female stress in free-ranging baboons. Anim Behav. 69:1211-1221. Bercovitch FB, Clarke AS. 1995. Dominance rank, cortisol concentrations, and reproductive maturation in male rhesus macaques. Physiol Behav. 58:215-221. Bernstein IS, Cooper MA. 1999. Do minance in Assamese macaques ( Macaca assamensis). Am J Primatol. 48:283-289. Bertram BCR. 1978. Living in groups: Predat ors and prey. In: Krebs JR, Davies NB, editors. Behavioral ecology: An evolut ionary approach. Oxford: Blackwell Scientific Publications. p. 64-96. Boccia ML, Reite M, Laudenslager M. 1989. On the physiology of grooming in a pigtail macaque. Physiol Behav. 45:667-670. Boinski S. 1994. Affiliation patterns among male Costa Rican squirrel monkeys. Behaviour. 130: 191-209. Boinski S. 1999. The social organizations of squirrel monkeys: Implications for ecological models of social evol ution. Evol Anthropol. 8:101-112. Boinski S, Gross TS, Davis JK. 1999a. Terrest rial predator alarm vocalizations are a valid monitor of stress in captive brown capuchins ( Cebus apella ). Zoo Biol. 18:295-312. Boinski S, Kauffman L, Westoll A, Stickler CM, Cropp S, Ehmke E. 2003. Are vigilance, risk from avian predators and group size consequences of habitat structure? A comparison of three species of squirrel monkey (Saimiri oerstedii, S. boliviensis, and S. sciureus ). Behaviour. 140:1421-1467. Boinski S, Sughrue K, Selvaggi L, Quatrone R, Henry M, Cropp S. 2002. An expanded test of the ecological model of primate social evolutio n: Competitive regimes and female bonding in three species of squirrel monkeys (Saimiri oerstedii, S. boliviensis, and S. sciureus ). Behaviour. 139: 227-261.
232 Boinski S, Swing SP, Gross TS, Davis JK. 1999b. Environmental enrichment of brown capuchins ( Cebus apella ): Behavioral and plasma and fecal cortisol measures of effectiveness. Am J Primatol. 48:49-68. Boonstra R, Hik D, Singleton GR, Tinniko v A. 1998. The impact of predator-induced stress on the snowshoe hare cycl e. Ecol Monogr. 68:371-394. Bradbury JW, Vehrencamp SL. 1976. Social or ganization and foraging in emballonurid bats. 2. Model for determination of group size. Behav Ecol Sociobiol. 1:383-404. Brown L, Amadon D. 1968. Eagles hawks, and falcons of the wo rld, Vol. 2. New York (NY): McGraw-Hill. Brunton PJ, Russell JA. 2003. Hypothalamic-pit uitary-adrenal responses to centrally administered orexin-A are suppressed in pregnant rats. J Neuroendocrinol. 15:633-637. Brunton PJ, Russell JA, Douglas AJ. 2008. Adaptive responses of the maternal hypothalamic-pituitary-adrenal axis during pregnancy and lactation. J Neuroendocrinol. 20:764-776. Byrne RW, Whiten A, Henzi SP. 1990. Soci al relationships of mountain baboons: Leadership and affiliation in a non-fem ale-bonded monkey. Am J Primatol. 20:313-329. Cameron EZ, Setsaas TH, Linklater WL. 2009. Social bonds between unrelated females increase reproductive success in feral horses. Proc Natl Acad Sci. 106: 1385013853. Campbell CJ, Shideler SF, Todd HF, Lasley BL. 2001. Fecal analysis of ovarian cycles in female black-handed spider monkeys ( Ateles geoffroyi ). Am J Primatol. 54:7989. Carnegie SD, Fedigan LM, Ziegler TE. 2005. B ehavioral indicators of ovarian phase in white-faced capuchins ( Cebus capucinus). Am J Primatol. 67:51-68. Carosi M, Heistermann M, Visalberghi E. 1999. Display of proceptive behaviors in relation to urinary and fecal progestin le vels over the ovarian cycle in female tufted capuchin monkeys. Horm Behav. 36:252-265. Carosi M, Visalberghi E. 2002. Analysis of tufted capuchin (Cebus apella ) courtship and sexual behavior repertoir e: Changes throughout the female cycle and female interindividual differences. Am J Phys Anthropol. 118:11-24. Carter CS. 1998. Neuroendocrine perspecti ves on social attachment and love. Psychoneuroendocrinology. 23:779-818.
233 Carter CS, Altemus M, Chrousos GP. 2001. Neuroendocrine and emotional changes in the post-partum period. Pr og Brain Res. 133:241-249. Carter CS, DeVries AC, Ge tz LL. 1995. Physiological substrates of mammalian monogamy: The prairie vole model. N eurosci Biobehav Rev. 19:303-314. Cavigelli SA. 1999. Behavioural patterns asso ciated with faecal cortisol levels in free-ranging female ring-tailed lemurs, Lemur catta Anim Behav. 57:935-944. Cavigelli SA, Dubovick T, Levash W, Jolly A, Pitts A. 2003. Female dominance status and faecal corticoids in a cooperative breeder with low reproductive skew: Ringtailed lemurs (Lemur catta ). Horm Behav. 43:166-179. Chapman CA, Chapman L. 2000. Determinant s of group size in primates: The importance of travel costs. In: Boinski S, Garber, PA, editors. On the move: How and why animals travel in groups. Chicago: University of Chicago Press. p. 2442. Chapman CA, Wasserman MD, Gillespie TR, Spei rs ML, Lawes MJ, Saj TL, Ziegler TE. 2006. Do food availability, parasitism, and stress have synergistic effects on red colobus populations living in forest fragments? Am J Phys Anthropol. 131:525534. Charmandari E, Tsigos C, Chrousos G. 2005. Endocrinology of the stress response. Annu Rev Physiol. 67:259-284. Cheney DL. 1992. Intragroup cohesion and intergroup hostility: The relation between grooming distributions and intergroup competition among female primates. Behav Ecol. 3:334-345. Cheney DL, Seyfarth RM. 1987. The influence of intergroup competition on the survival and reproduction of female vervet monk eys. Behav Ecol Sociobiol. 21:375-386. Cho MM, DeVries C, Williams JR, Carter CS. 1999. The effects of oxytocin and vasopressin on partner preferences in male and female prairie voles (Microtus ochrogaster) Behav Neurosci. 113: 1071-1079. Clutton-Brock TH, Harvey PH. 1977. Primate ecology and social organization. J Zool. 183:1-39. Coe CL, Levine S. 1995. Diur nal and annual variation of adrenocortical activity in the squirrel monkey. Am J Primatol. 35:283-292.
234 Coe CL, Murai JT, Wiener SG, Levine S, Siiteri PK. 1986. Rapid cortisol and corticosteroid-binding globulin responses during pregnancy and after estrogen administration in the squirrel monkey. Endocrinology. 118:435-440. Cook CJ. 1997. Oxytocin and prolactin suppre ss cortisol responses to acute stress in both lactating and non-lactating sheep. J Dairy Res. 64:327-339. Cowlishaw G. 1995. Behavioral patterns in baboon group encounters: The role of resource competition and male reproduc tive strategies. Behaviour. 132:75-86. Cowlishaw G. 1997a. Alarm calli ng and implications for risk perception in a desert baboon population. Et hology. 103:384-394. Cowlishaw G. 1997b. Trade-offs between foraging and predation risk determine habitat use in a desert baboon population. Anim Behav. 53: 667-686. Cowlishaw G, Dunbar RIM. 1991. Dominanc e rank and mating success in male primates. Anim Behav. 41:1045-1056. Creel SF, 2005. Dominance, aggression, and gluc ocorticoid levels in social carnivores. J Mammal. 86:255-264. Crockford C, Wittig RA, Whitten PL, Seyfarth RA, Cheney DL. 2008. Social stressors and coping mechanisms in wild female baboons ( Papio hamadryas ursinus ). Horm Behav. 53:254-265. Curley JP, Keverne EB. 2005. Genes, brains and mammalian social bonds. Trends Ecol Evol. 20: 561-567. Curry MR, Eady PE, Mills DS. 2007. Reflec tions on mare behavior: Social and sexual perspectives. J Vet Behav. 2: 149-157. de Waal FBM. 1986. The integr ation of dominance and soci al bonding in primates. Q Rev Biol. 61: 459-479. de Waal FBM, Luttrell LM. 1986. The similarity principl e underlying social bonding among female rhesus monkeys. Folia Primatol. 46:215-234. de Ruiter J. 1986. The influence of gr oup size on predator scanning and foraging behavior of wedge-capped capuchin monkeys (Cebus olivaceus). Behaviour. 98:240-258. Defler TR. 1982. A comparison of intergroup behavior in Cebus albifrons and C. apella Primates. 23:385-392.
235 Di Bitetti MS. 1997. Evidence for an important soci al role of allogrooming in a platyrrhine primate. Anim Behav. 54:199-211. Di Bitetti MS. 2001. Home-range use by the tufted capuchin monkey ( Cebus apella nigritus ) in a subtropical rainforest of Argentina. J Zool. 253:33-45. Di Bitetti MS, Janson CH. 2001. Reproductive socioecology of tufted capuchins ( Cebus apella nigritus ) in Northeastern Argentina. Int J Primatol. 22:127-142. Donner N, Bredewold R, Maloumby R, N eumann ID. 2007. Chronic intracerebral prolactin attenuates neuronal stress circuitr ies in virgin rats. Eur J Neurosci. 25:1804-1814. Drea CM, Frank LG. 2003. The so cial complexity of spotted hyenas. In: de Waal FBM, Tyack PL, editors. Animal social comple xity. Cambridge: Harvard University Press. p.121-148. Dunbar RIM. 1987. Habitat quality, populat ion dynamics and group composition in colobus monkeys ( Colobus guereza ). Int J Primatol. 8:299-329. Dunbar RIM. 1988. Primate social systems. Ne w York (NY): Cornell University Press. Dunbar RIM. 1991. Functional significance of social grooming in primates. Folia Primatol. 57: 121-131. Dunbar RIM. 1996. Grooming, gossip and the evolution of language. London (England): Faber & Faber. Dunbar RIM. 2009. The social brain hypothesis and its implications fo r social evolution. Ann Hum Biol. 36: 562-572. East ML, Hofer H, Wickler W. 1993. The er ect penis as a flag of submission in a female-dominated society: Greetings in Serengeti spotted hyenas. Behav Ecol Sociobiol. 33: 335-370. Eckstein RA, Hart BL. 2000. Grooming and control of fleas in cats. Appl Anim Behav Sci. 68: 141-150. Eisenberg JF, Muckenheim NA, Rudran R. 1972. Relation between ecology and social structure in primates Science. 176:863-874. Elgar MA. 1989. Predator vigilance and group si ze in mammals and birds: A critical review of the empirical evidence. Bi ol Rev Camb Philos Soc. 64:13-33. Emlen ST, Oring LW. 1977. Ecology, sexual se lection, and evolution of mating systems. Science. 197:215-223.
236 Engh AL, Beehner JC, Bergman TJ, Whitten PL, Hoffmeier RR, Seyfarth RM, Cheney DL. 2006a. Female hierarchy instability, male immigration and infanticide increase glucocorticoid levels in female chacma baboons. Anim Behav. 71:12271237. Engh AL, Beehner JC, Bergman TJ, Whitten PL, Hoffmeier RR, Seyfarth RM, Cheney DL. 2006b. Behavioural and hormonal responses to predation in female chacma baboons ( Papio hamadryas ursinus ). Proc Biol Sci. 273:707-712. Fedigan L. 1993. Sex differences and inte rsexual relations in adult white-faced capuchins ( Cebus capucinus). Int J Primatol. 14:853-877. Fedurek P, Dunbar RIM. 2009. What does mutual grooming tell us about why chimpanzees groom? Et hology. 115:566-575. Feh C, de Mazieres J. 1993. Grooming at a pr eferred site reduces heart rate in horses. Anim Behav. 46: 1191-1194. Ferreira RG, Izar P, Lee PC. 200 6. Exchange, affiliation, and protective interventions in semifree-ranging brown capuchin monkeys ( Cebus apella ). Am J Primatol. 68:765-776. Foley CAH, Papageorge S, Wasser SK. 2001. Noninvasive stress and reproductive measures of social and ecological pre ssures in free-ranging African elephants. Conserv Biol. 15:1134-1142. Fragaszy DM, Visalberghi E, Fedigan LM. 2004. The complete capuchin: The biology of the genus Cebus Cambridge (England): Cambridge University Press. Franz C. 1998. Allogrooming behavior and grooming site preferences in captive bonobos ( Pan paniscus ): Association with female dominance. Int J Primatol. 20:525-546. Freese CH, Oppenheimer JR. 1981. The capuchin monkey, genus Cebus In: Coimbra-Filho AF, Mittermeier RA, edi tors. Ecology and behavior of neotropical primates, Vol 1. Rio de Janeiro: Academ ia Brasileira de Ci encias. p. 331-390. Ganas J, Robbins MM. 2005. Ranging be havior of the mountain gorillas ( Gorilla beringei beringei ) in Bwindi Impenetrable National Park, Uganda: A test of the ecological constraints model. Behav Ecol Sociobiol. 58:277-288. Gesquiere LR, Khan M, Shek L, Wango TL, Wango EO, Alberts SC, Altmann J. 2008. Coping with a challenging environment: Ef fects of seasonal variability and reproductive status on glucocorticoid concentrations of female baboons ( Papio cynocephalus ). Horm Behav. 54:410-416.
237 Getz LL, Carter CS. 1996. Prairie-vo le relationships. Am Sci. 84: 56-62. Gil-da-Costa R, Palleroni A, Hauser MD, Touchton J, Patrick Ke lley J. 2003. Rapid acquisition of an alarm response by a neot ropical primate to a newly introduced avian predator. Proc Biol Sci. 270:605-610. Glynn LM, Schetter CD, Wadhwa PD, Sand man CA. 2004. Pregnancy affects appraisal of negative life events. J Psychosom Res. 56:47-52. Goland RS, Josak S, Conwell I. 1994. Plac ental corticotropin-releasing hormone and the hypercortisolism of pregnancy. Am J Obstet Gynecol. 171:1287-1291. Gould L, Ziegler TE, Wittwer DJ. 2005. Effects of reproductive and social variables on fecal glucocorticoid levels in a samp le of adult male ring-tailed lemurs ( Lemur catta ) at the Beza Mahafaly Reserve, M adagascar. Am J Primatol. 67:5-23. Goymann W, Mostl E, Van't Hof T, East ML, Hofer H. 1999. Noninvasive fecal monitoring of glucocorticoids in spotted hyenas, Crocuta crocuta Gen Comp Endocrinol. 114:340-348. Gust DA, Gordon TP, Brodie AR, McClure HM. 1994. Effect of a preferred companion in modulating stress in adult female rhesus monkeys. Physiol Behav. 53:681-684. Gust DA, Gordon TP, Hambright MK, Wils on ME. 1993. Relationship between social factors and pituitary-adrenocortical acti vity in female rhesus monkeys ( Macaca mulatta ). Horm Behav. 27:318-331. Haller J, Fuchs E, Halasz J, Makara GB. 1999. Defeat is a major stressor in males while social instability is stressful mainly in females: Towards the development of a social stress model in female rats. Brain Res Bull. 50:33-39. Hart BL, Hart LA. 1992. Reci procal grooming in impala, Aepyceros melampus Anim Behav. 44:1073-1083. Hart BL, Pryor PA. 2004. Developmental and hair-coat determinants of grooming behaviour in goats and sheep. Anim Behav. 67: 11-19. Hawlena H, Bashary D, Abramsky Z, Khok hlova IS, Krasnov BR. 2008. Programmed versus stimulus-driven antiparasitic gr ooming in a desert rodent. Behav Ecol. 19: 929-935. Heinrichs M, Baumgartner T, Kirschbaum C, Ehlert U. 2003. Social support and oxytocin interact to suppress cortisol and subjective responses to psychosocial stress. Biol Psychiatry. 54:1389-1398.
238 Heinrichs M, Meinlschmidt G, Hellh ammer DH. 2001. Effects of suckling on hypothalamic-pituitary-adrenal axis responses to psychosocial stress in postpartum lactating women. J Clin Endocrinol Metab. 86:4798-4804. Heitor F, Oom MM, Vicente L. 2006. Social relationships in a herd of Sorraia horses. Part II. Factors affecting affiliative relationships and sexual behaviors. Behav Processes. 73: 231-239. Hemelrijk CK. 1999. An individual-oriented model of the emergence of despotic and egalitarian socities. Proc Biol Sci. 266:361-369. Hemelrijk CK, Luteijn M. 1998. Philopatry, male presence and grooming reciprocation among female primates: A comparative per spective. Behav Ecol Sociobiol. 42: 207-215. Hennessy MB, Hornschuh G, Kaiser S, Sachser N. 2006. Cortisol responses and social buffering: A study throughout the life span. Horm Behav. 49:383-390. Hennessy MB, Zate R, Maken DS. 2008. Social buffering of the cortisol response of adult female guinea pigs. Physiol Behav. 93:883-888. Henzi SP, Barrett L. 1999. The value of groom ing to female primates. Primates. 40:4759. Henzi SP, Barrett L. 2002. Infants as a commodity in a baboon market. Anim Behav. 63:915-921. Henzi SP, Barrett L, Gaynor D, Greeff J, Weingrill T, Hill RA. 2003. Effect of resource competition on the long-term alloca tion of grooming by female baboons: Evaluating Seyfarth's model. Anim Behav. 66:931-938. Henzi SP, Lusseau D, Weingrill T, van Schaik CP, Barrett L. 2009. Cyclicity in the structure of female baboon social netwo rks. Behav Ecol Sociobiol. 63: 10151021. Henzi SP, Lycett JE, Weingrill A, Piper SE. 2000. Social bonds and the coherence of mountain baboon troops. Behaviour. 137: 663-680. Hill WCO. 1960. Primates: Compar ative anatomy and taxonomy. IV. Cebidae. Part A. Edinburgh (Scotland): University of Edinburgh Press. Hirsch BT. 2002. Social monitoring and vigil ance behavior in brown capuchin monkeys ( Cebus apella ). Behav Ecol Sociobiol. 52: 458-464.
239 Holekamp KE, Cooper SM, Katona CI, Berry NA, Frank LG, Smale L. 1997. Patterns of association among female spotted hyenas ( Crocuta crocuta). J Mammal. 78: 5564. Insel TR, Young LJ. 2001. The neurobiology of attachment. Nat Rev Neurosci. 2:129136. Isbell LA. 1994. Predation on primates: Ecological patterns and evolutionary consequences. Evol Anthropol. 3: 61-71. Isbell LA, Young TP. 1993. Social and ecologic al influences on activity budgets of vervet monkeys, and their implications for group living. Behav Ecol Sociobiol. 32:377-385. Isvaran K. 2007. Intraspecific variation in group size in t he blackbuck antelope: the roles of habitat structure and forage at diffe rent spatial scales. Oecologica. 154:435444. Izar P. 2004. Female social relationships of Cebus apella nigritus in a southeastern Atlantic Forest: An analysis through ec ological models of primate social evolution. Behaviour. 141:71-99. Izawa K. 1980. Social behavior of the wild black-capped capuchin. Primates. 21:443467. Jack KM. 2003. Explaining variation in affilia tive relationships among male white-faced capuchins ( Cebus capucinus). Folia Primatol. 74: 1-16. Janson C. 1984. Female choice and mating system of the brown capuchin monkey Cebus apella (Primates, Cebidae). J Comp Ethol. 65:177-200. Janson C. 1985. Aggressive competition and individual food consumption in wild brown capuchin monkeys ( Cebus apella ). Behav Ecol Sociobiol. 18:125-138. Janson C. 1996. Toward an experimental so cioecology of primates. In: Norconk MA, Rosenberger AL, Garber PA, editors. Adaptive radiations of neotropical primates. New York: Plenum Press. p. 309-325. Janson CH. 1988a. Food competition in brown capuchin monkeys ( Cebus apella): Quantitative effects of group size and tree productivity. Behaviour. 105:53-76. Janson CH. 1988b. Intra-specific food compet ition and primate soci al structure: A synthesis. Behaviour. 105:1-17. Janson CH. 1990a. Ecological consequences of individual spatial choice in foraging groups of brown capuchin monkeys, Cebus apella Anim Behav. 40:922-934.
240 Janson CH. 1990b. Social correlates of individual spatial choice in foraging groups of brown capuchin monkeys, Cebus apella. Anim Behav. 40:910-921. Janson CH, Boinski S. 1992. Morphological and behavioral adaptations for foraging in generalist primates: The case of the Cebines. Am J Phys Anthropol. 88:483-498. Janson CH, Goldsmith ML. 1995. Predicting group size in primates: foraging costs and predation risks. Behav Ecol. 6:326-336. Kaiser S, Kirtzeck M, Hornschuh G, Sachser N. 2003. Sex-specific difference in social support: A study in female guinea pigs. Physiol Behav. 79:297-303. Kammerer M, Adams D, von Castelberg B, Glover V. 2002. Pregnant women become insensitive to cold stress. BMC Pregnancy Childbirth. 2:8. Kapsalis E, Berman CM. 1996. Models of aff iliative relationships among free-ranging rhesus monkeys ( Macaca mulatta ). 2. Testing predictions for three hypothesized organizing principles. Behaviour. 133:1235-1263. Keeling LJ, Estevez I, Newberry RC, Correia MG. 2003. Production-related traits of layers reared in different sized flocks: the concept of problematic intermediate group sizes. Poult Sci. 82:1393-1396. Keverne EB, Martensz ND, Tuite B. 1989. Beta-endorphin concentrations in cerebrospinal fluid of monkeys are influenced by grooming relationships. Psychoneuroendocrinology. 14:155-161. Kinzey WG, Wright PC. 1982. Groomi ng behavior in the titi monkey ( Callicebus torquatus ). Am J Primatol. 3:267-275. Kirschbaum C, Klauer T, Filipp SH, Hellhammer DH. 1995. Sex-specific effects of social support on cortisol and subjective res ponse to acute psychological stress. Psychosom Med. 57:23-31. Koenig A. 2000. Competitiv e regimes in forest-dwelling Hanuman langur females ( Semnopithecus entellus ). Behav Ecol Sociobiol. 48:93-109. Kohari D, Sato S, Nakai Y. 2009. Does t he maternal grooming of cattle clean bacteria from the coat of calves? Behav Processes. 80: 202-204. Kondo S, Sekine J, Okubo M, Asahida Y. 1989. The effect of group size and space allowance on the agonistic and spacing behavi our in cattle. Appl Anim Behav Sci. 24:127-135.
241 Krause J, Ruxton GD. 2002. Li ving in groups. Oxford (E ngland): Oxford University Press. Krebs CJ, Boutin S, Boonstra R, Sinclair ARE, Smith JNM, Dale MRT, Martin K, Turkington R. 1995. Impact of food and predation on the snowshoe hare cycle. Science. 269:1112-1115. Kutsukake N. 2009. Complexity, dynamics and diversity of sociality in group-living mammals. Ecol Res. 24: 521-531. Kutsukake N, Clutton-Brock TH. 2006. Social f unctions of allogrooming in cooperatively breeding meerkats. Anim Behav. 72: 1059-1068. Lahoz MM, Nagle CA, Porta M, Farinati Z, Manzur TD. 2007. Cortisol response and ovarian hormones in juvenile and cycling female Cebus monkeys: Effect of stress and Dexamethasone. Am J Primatol. 69:551-561. Leca JB, Fornasieri I, Petit O. 2002. Aggression and reconciliation in Cebus capucinus. Int J Primatol. 23: 979-998. Legros JJ. 2001. Inhibitory effe ct of oxytocin on corticotrope functions in humans: Are vasopressin and oxytocin ying-yang neurohormones? Psychoneuroendocrinology. 26:649-655. Lehmann J, Korstjens AH, Dunbar RIM. 2007. Group size, grooming and social cohesion in primates. Anim Behav. 74:1617-1629. Linn GS, Mase D, Lafrancois D, O'Keeffe RT Lifshitz K. 1995. Social and menstrual cycle phase influences on the behavior of group-housed Cebus apella Am J Primatol. 35:41-57. Lockwood CJ, Radunovic N, Nastic D, Petk ovic S, Aigner S, Berkowitz GS. 1996. Corticotropin-releasing hormone and relat ed pituitary-adrenal axis hormones in fetal and maternal blood during the se cond half of pregnancy. J Perinat Med. 24:243-251. Lott DF. 1984. Intraspecific variation in the social systems of wild vertebrates. Behaviour. 88: 266-325. Lyimo ZC, Nielen M, Ouweltjes W, Kr uip TAM, van Eerdenburg FJCM. 2000. Relationship among estradiol, cortisol and in tensity of estrous behavior in dairy cattle. Theriogenology. 53:1783-1795. Lynch JW, Ziegler TE, Stri er KB. 2002. Individual and s easonal variation in fecal testosterone and cortisol levels of wild male tufted capuchin monkeys, Cebus apella nigritus Horm Behav. 41:275-287.
242 Madden JR, Clutton-Brock TH. 2009. M anipulating grooming by decreasing ectoparasite load causes unpredicted changes in antagonism. Proc Biol Sci. 276:1263-1268. Maestripieri D, Lindell SG, Ayala A, Go ld PW, Higley JD. 2005. Neurobiological characteristics of rhesus macaque abusive mothers and their relation to social and maternal behavior. Neuro sci Biobehav Rev. 29:51-57. Manogue KR, Leshner AI, Candland DK. 1975. Dominance status and adrenocortical reactivity to stress in squirrel monkeys (Saimiri sciureus ). Primates. 16:457-463. Manson JH, Rose LM, Perry S, Gros-Loui s J. 1999. Dynamics of female-female relationships in wild Cebus capucinus : Data from two Costa Rican sites. Int J Primatol. 20:679-706. Martel FL, Nevison CM, Ra yment FD, Simpson MJA, Keverne EB. 1993. Opioid receptor blockade reduces maternal a ffect and social grooming in rhesus monkeys. Psychoneuroimmunology. 18:307-321. Mateo JM. 2007. Ecological and hormonal correlates of antipredator behavior in adult Belding's ground squirrels ( Spermophilus beldingi ). Behav Ecol. 62:37-49. Matheson MD, Bernstein IS. 2000. Grooming, social bonding and agonistic aiding in rhesus monkeys. Am J Primatol. 51:177-186. McCarthy MM, Altemus M. 1997. Central nervous system actions of oxytocin and modulation of behavior in humans. Mol Med Today. 3:269-275. McCormick CM, Smythe JW, Sharma S, M eaney MJ. 1995. Sex-specific effects of prenatal stress on hypothalamic-pituitary-adrenal responses to stress and brain glucocorticoid receptor density in adult rats. Dev Brain Res. 84:55-61. Mendoza SP, Lyons DM, Saltzman W. 1991. So ciophysiology of squirrel monkeys. Am J Primatol. 23:37-54. Mitchell CL, Boinski S, van Schaik CP. 1991. Competitive regimes and female bonding in two species of squirrel monkeys ( Saimiri oerstedi and S. sciureus). Behav Ecol Sociobiol. 28:55-60. Mooring MS, McKenzie AA, Hart BL. 1996. Gr ooming in impala: Role of oral grooming in removal of ticks and effects of ticks in increasing grooming rate. Physiol Behav. 59: 965-971. Mooring MS, Samuel WM. 1998a. Tick defense strategies in bison: The role of grooming and hair coat. Behaviour. 135: 693-718.
243 Mooring MS, Samuel WM. 1998b. The biolog ical basis of grooming in moose: Programmed versus stimulus-driven grooming. Anim Behav. 56: 1561-1570. Moura A. 2007. Primate group size and abundance in the caatinga dry forest, northeastern Brazil. In t J Primatol. 28:1279-1297. Moynihan MA. 1964. Some behavioral patterns of platyrrhine monkeys. 1. The night monkey (Atous trivirgatus ). Smith Misc Collect. 146: 1-84. Muroyama Y. 1994. Exchange of grooming for allomothering in female patas monkeys. Behaviour. 128:103-119. Nakamichi M, Shizawa Y. 2003. Distribution of groomin g among adult females in a large, free-ranging group of Japanese macaques. Int J Primatol. 24: 607-625. Neumann ID, Torner L, Wigger A. 2000a. Brai n oxytocin: Differential inhibition of neuroendocrine stress responses and anxie ty-related behaviour in virgin, pregnant and lactating rats. Neuroscience. 95:567-575. Neumann ID, Wigger A, Torner L, Holsboer F, Landgraf R. 2000b. Brain oxytocin inhibits basal and stress-induced activity of the hypothalamo-pituitary-adrenal axis in male and female rats: Partial ac tion within the paraventricular nucleus. J Neuroendocrinol. 12:235-243. Norconk MA, Sussman RW, Phillips-Conroy J. 1996. Primates of Guayana Shield forests: Venezuela and the Guianas. In: Norconk MA, Rosenberger AL, Garber PA, editors. Adaptive radiations of neot ropical primtaes. New York: Plenum Press. p. 69-84. Norcross JL, Newman JD. 1999. Effects of separation and novelty on distress vocalizations and cortisol in the common marmoset ( Callithrix jacchus ). Am J Primatol. 47:209-222. O'Brien TG. 1991. Female-male social in teractions in wedge-capped capuchin monkeys: Benefits and costs of group living. Anim Behav. 41:555-567. O'Brien TG. 1993. Allogrooming behavior among adult female wedge-capped capuchin monkeys. Anim Behav. 46:499-510. O'Brien TG, Robinson JG. 1993. Stability of social relationships in female wedge-capped capuchin monkeys. In: Pereira ME, Fairbanks LA, editors. Juvenile primates: Life history, de velopment, and behavior. Oxford: Oxford University Press. p. 197-210.
244 Olsson IAS, Westlund K. 2007. More than numbers matter: The effect of social factors on behaviour and welfare of laboratory rodents and nonhuman primates. Appl Anim Behav Sci. 103:229-254. Palme R, Rettenbacher S, Touma C, El-Bahr SM, Mostl E. 2005. Stress hormones in mammals and birds: Comparative aspects regarding metabolism, excretion, and noninvasive measurement in fecal samples. Ann N Y Acad Sci. 1040:162-171. Palombit RA, Seyfarth RM, Cheney DL. 1997. The adaptive value of 'friendships' to female baboons: Experimental and observa tional evidence. Anim Behav. 54: 599-614. Parr LA, Matheson MD, Bernstein IS, de Waal FBM. 1997. Gr ooming down the hierarchy: Allogrooming in c aptive brown capuchin monkeys, Cebus apella. Anim Behav. 54:361-367. Payne HFP, Lawes MJ, Henzi SP. 2003. Competition and the exchange of grooming among female samango monkeys ( Cercopithecus mitis erythrarchus ). Behaviour. 140:453-471. Peel AJ, Vogelnest L, Finniga n M, Grossfeldt L, O'Brien JK. 2005. Non-invasive fecal hormone analysis and behavioral observations for monitoring stress responses in captive western lowland gorillas ( Gorilla gorilla gorilla). Zoo Biol. 24:431-445. Pepe GJ, Albrecht ED. 1995. Actions of plac ental and fetal adrenal-steroid hormones in primate pregnancy. En docr Rev. 16:608-648. Perry S. 1996. Female-female social rela tionships in wild white-faced capuchin monkeys, Cebus capucinus. Am J Primatol. 40:167-182. Perry S, Manson JH, Muniz L, Gros-Louis J, Vigilant L. 2008. Kin-biased social behaviour in wild adult female white-faced capuchins, Cebus capucinus Anim Behav. 76:187-199. Pollard KA, Blumstein DT. 2008. Time alloca tion and the evolution of group size. Anim Behav. 76: 1683-1699. Price EE, Stoinski TS. 2007. Group size: Dete rminants in the wild and implications for the captive housing of wild mammals in zoos. Appl Anim Behav Sci. 103: 255264. Pride E. 2005a. Optimal group size and s easonal stress in ring-tailed lemurs ( Lemur catta ). Behav Ecol. 16:550-560. Pride RE. 2005b. Foraging success, agoni sm, and predator alarms: Behavioral predictors of cortisol in Lemur catta Int J Primatol. 26:295-319.
245 Proctor CJ, Broom M, and Ruxton GD. 2001. Modelling antipredator vigilance and flight response in group foragers when warning signals are ambiguous. J Theor Biol. 211: 409-417. Pulliam HR. 1973. On the advantages of flocking. J Theor Biol. 38:419-422. Pulliam HR, Caraco T. 1984. Living in groups : Is there an optimal group size? In: Krebs JR, Davies JB, editors. Behavioural ecology: An evolutionary approach. Sunderland: Sinauer Associates. p. 122-147. Quenette PY. 1990. Functions of vigilance in mammals: A review. Acta Oecol. 11:801818. Ram S, Venkatachalam S, Sinha A. 2003. Changing social strategies of wild female bonnet macaques during natural foraging an d on provisioning. Curr Sci. 84:780790. Ramos-Fernandez G, Boyer D, Aureli F, Vick LG. 2009. Association networks in spider monkeys ( Ateles geoffroyi ). Behav Ecol Sociobiol. 63: 999-1013. Reeder DM, Kramer KM. 2005. Stress in fr ee-ranging mammals: Integrating physiology, ecology, and natural histor y. J Mammal. 86:225-235. Reichart H. 1993. Management plan for Rale ighvallen Nature Preserve, 1993-1998. Paramaribo (Suriname): Suriname Nature Conservation Foundation. Rieucau G, Martin JGA. 2008. Many eyes or many ewes: vigilance tactics in female bighorn sheep Ovis canadensis vary according to reproductive status. Oikos. 117:501-506. Robinson JG. 1988. Group size in wedge-capped capuchin monkeys ( Cebus olivaceus) and the reproductive success of males and females. Behav Ecol Sociobiol. 23:187-197. Rodenburg TB, Koene P. 2007. The impact of group size on damaging behaviours, aggression, fear and stress in farm anima ls. Appl Anim Behav Sci. 103:205-214. Rosal MC, King J, Ma Y, Reed GW. 2004. Stre ss, social support, and cortisol: Inverse associations? Behav Med. 30:11-21. Rubenstein DI. 1981. Behavioural ecology of island feral horses. Equine Vet J. 13:2734.
246 Rukstalis M, French JA. 2005. Vocal bufferi ng of the stress response: Exposure to conspecific vocalizations moderates urinary cortisol excretion in isolated marmosets. Horm Behav. 47:1-7. Sachser N, Durschlag M, Hirzel D. 1998. So cial relationships and the management of stress. Psychoneuroendoc rinology. 23:891-904. Sade DS. 1972. Sociometrics of Macaca mulatta I. Linkages and cliques in grooming matrices. Folia Primatol. 18:196-223. Saltzman W, Mendoza SP, Mason WA. 1991. Sociophysiology of relationships in squirrel monkeys. 1. Formation of fe male dyads. Physiol Behav. 50: 271-280. Saltzman W, Schultz-Darken NJ, Wegner FH, Wittwer DJ, Abbott DH. 1998. Suppression of cortisol levels in s ubordinate female marmosets: Reproductive and social contributions. Horm Behav. 33:58-74. Sambrook TD, Whiten A, Strum SC. 1995. Prio rity of access and grooming patterns of females in a large and a small group of olive baboons. Anim Behav. 50: 16671682. Sanchez-Andrade G, Kendrick KM. 2009. The main olfactory system and social learning in mammals. Behav Brain Res. 200: 323-335. Sapolsky RM. 1986. Endocrine and behavioral correlates of drought in wild olive baboons ( Papio anubis ). Am J Primatol. 11:217-227. Sapolsky RM. 2005. The influence of social hierarchy on primate health. Science. 308:648-652. Sapolsky RM, Romero LM, Munck AU. 2000. How do glucocorticoids influence stress responses? Integrating permissive, suppr essive, stimulatory, and preparative actions. Endocr Rev. 21:55-89. Schaffner CM, Aureli F. 2005. Embraces and gr ooming in captive spider monkeys. Int J Primatol. 26: 1093-1106. Scheuerlein A, Van't Hof TJ, Gwinner E. 2001. Predators as stressors? Physiological and reproductive consequences of predation risk in tropical stonechats ( Saxicola torquata axillaris ). Proc Biol Sci. 268:1575-1582. Schino G. 2001. Grooming, competition and social rank among female primates: A meta-analysis. Anim Behav. 62:265-271. Schino G. 2007. Grooming and agonistic support: A meta-analysis of primate reciprocal altruism. Behav Ecol. 18:115-120.
247 Schino G, Aureli F. 2008. Trade-offs in pr imate grooming reciprocation: Testing behavioural flexibility and correlated ev olution. Biol J Linn Soc. 95:439-446. Schino G, Di Giuseppe F, Visalberghi E. 2009. Grooming, rank, and agonistic support in tufted capuchin monkeys. Am J Primatol. 71:101-105. Schino G, Scucchi S, Maestripieri D, Tur illazzi PG. 1988. Allogrooming as a tension reduction mechanism: A behavioral approac h. Am J Primatol. 16:43-50. Shanks N, Windle RJ, Perks P, Wood S, Ingram CD, Lightman SL. 1999. The hypothalamic-pituitary-adrenal axis res ponse to endotoxin is attenuated during lactation. J Neuroendocrinol. 11:857-865. Shively CA, Register TC, Friedman DP, Morgan TM, Thompson J, Lanier T. 2005. Social stress-associated depression in adult female cynomolgus monkeys ( Macaca fascicularis ). Biol Psychol. 69:67-84. Shutt K, MacLarnon A, Heistermann M, Semple S. 2007. Grooming in Barbary macaques: Better to give than to receive? Biol Lett. 3:231-233. Seyfarth RM. 1977. Model of social grooming among adult female monkeys. J Theor Biol. 65:671-698. Seyfarth RM, Cheney DL. 1984. Grooming, allianc es and reciprocal altruism in vervet monkeys. Nature. 308:541-543. Shively CA, Register TC, Friedman DP, Morgan TM, Thompson J, Lanier T. 2005. Social stress-associated depression in adult female cynomolgus monkeys ( Macaca fascicularis ). Biol Psychol. 69:67-84. Shultz S, Dunbar RIM. 2007. The evolution of the social brain: Anthropoid primates contrast with other vertebrates Proc Biol Sci. 274: 2429-2436. Shumaker RW and Beck BB. 2003. Primates in question: The Smithsonian answer book. Washington (DC): Sm ithsonian Institution. Silk J, 1982. Altr uism among female Macaca radiata : Explanations and analysis patters of grooming and coalition fo rmation. Behaviour. 79:162-168. Silk JB. 2002. Using the F-word in primatology. Behaviour. 139: 421-446. Silk JB. 2007a. The adaptive value of social ity in mammalian groups. Philos Trans R Soc Lon B Biol Sci. 362:539-559.
248 Silk JB. 2007b. Social components of fi tness in primate gr oups. Science. 317:13471351. Silk JB, Alberts SC, Altmann J. 2003. Soci al bonds of female baboons enhance infant survival. Science. 302:1231-1234. Silk JB, Alberts SC, Altmann J. 2004. Pattern s of coalition formati on by adult female baboons in Amboseli, Kenya. Anim Behav. 67:573-582. Silk JB, Altmann J, Alberts SC. 2006a. Social relationships among adult female baboons ( Papio cynocephalus ) I. Variation in the strength of social bonds. Behav Ecol Sociobiol. 61:183-195. Silk JB, Alberts SC, Altmann J. 2006a. Social relationships among adult female baboons ( Papio cynocephalus ) II. Variation in the quality and stability of social bonds. Behav Ecol Sociobiol. 61:197-204. Silk JB, Seyfarth RM, Cheney DL. 1999. The structure of social relationships among female savanna baboons in Moremi Reserve, Botswana. Behaviour. 136:679703. Slater KY, Schaffner CM, Aureli F. 2007. Embraces for infant handling in spider monkeys: Evidence for a biological market? Anim Behav. 74: 455-461. Slattery DA, Neumann ID. 2008. No st ress please! Mechanisms of stress hyporesponsiveness of the maternal brain. The J Physiol. 586:377-385. Smith TF, French JA. 1997. Soci al and reproductive conditions modulate urinary cortisol excretion in black tufted-ear marmosets ( Callithrix leuhli ). Am J Primatol. 42:253267. Smith JE, Memenis SK, Hole kamp KE. 2007. Rank-related partner choice in the fission-fusion society of the spotted hyena ( Crocuta crocuta). Behav Ecol Sociobiol. 61: 753-765. Snaith TV, Chapman CA. 2007. Primate group si ze and interpreting sociolecological models: Do folivores really play by di fferent rules? Evol Anthropol. 16:94-106. Soto-Gamboa M, Villalon M, Bozinovic F. 2005. Social cues and hormone levels in male Octodon degus (Rodentia): A field test of the challenge hypothesis. Horm Behav. 47:311-318. Sousa MBC, Ziegler TE. 1998. Diurnal vari ation on the excretion patterns of fecal steroids in common marmoset ( Callithrix jacchus) females. Am J Primatol. 46:105-117.
249 Stammbach E. 1978. On social differentiati on in groups of captive female hamadryas baboons. Behaviour. 67:322-338. Stammbach E, Kummer H. 1982. Individual cont ributions to a dyadic interaction: An analysis of baboon grooming. Anim Behav. 30:964-971. Stavisky RC, Adams MR, Watson SL, K aplan JR. 2001. Dominance, cortisol, and behavior in small groups of female cynomolgus monkeys ( Macaca fascicularis ). Horm Behav. 39:232-238. Sterck EHM, Watts DP, van Schaik CP. 1997. The evolution of female social relationships in nonhuman primates. Behav Ecol Sociobiol. 41:291-309. Strier KB. 1992a. Faces in the forest: T he endangered muriqui monkeys of Brazil. New York (NY): Oxford University Press. Strier KB. 1992b. Causes and consequences of nonaggression in woolly spider monkeys. In: Silverberg J, Gray JP editors. Aggression and peacefulness in humans and other primates. New York: Ox ford University Press. p. 100-116. Strier KB. 1999. Why is female kin bonding so rare? Comparative sociality of neotropical primates. In: Lee PC, editor. Comparative primate socioecology. Cambridge: Cambridge University Press. p. 300-319. Strier KB, Carvalho DS, Bejar NO. 2000. Pre scription for peacefulness. In: Aureli F, de Waal FBM, editors. Natural conflict re solution. Los Angeles: University of California Press. p. 315-317. Strier KB, Dib LT, Figueira JEC. 2002. Social dynamics of male muriquis (Brachyteles arachnoids hypoxanthus ). Behaviour. 139: 315-342. Swedell L. 2002. Affiliation among females in wild hamadryas baboons ( Papio hamadryas hamadryas ). Int J Primatol. 23:1205-1226. Takahata Y, Suzuki S, Okayasu N, Sugiura H, Takahashi H, Yamagiwa J, Izawa K, Agetsuma N, Hill D, Saito C, Sato S, Tanaka T, Sprague D. 1998. Does troop size of wild Japanese macaques influence birth rate and infant mortality in the absence of predators? Primates. 39:245-251. Tamashiro KLK, Nguyen MMN, Sakai RR. 2 005. Social stress: From rodents to primates. Front Neuroendocrinol. 26:27-40. Taylor SE, Klein LC, Lewis BP, Gruenewal d TL, Gurung RAR, Updegraff JA. 2000. Biobehavioral responses to stress in females: Tend-and-befriend, not fight-or-flight. Psychol Rev. 107:411-429.
250 Terborgh J. 1983. Five New World primates: A study of comparative ecology. Princeton (NJ): Princeton University Press. Thierry B, Gauthier C, Pe ignot P. 1990. Social groom ing in Tonkean macaques ( Macaca tonkeana ). Int J Primatol. 11:357-375. Tilbrook AJ, Clarke IJ. 2006. Neuroendocri ne mechanisms of innate states of attenuated responsiveness of the hypothalam o-pituitary adrenal axis to stress. Front Neuroendocrinol. 27:285-307. Tilbrook AJ, Turner AL, Ibbott MD Clarke IJ. 2006. Activation of the hypothalamo-pituitary-adrenal axis by isolation and restraint stress during lactation in ewes: Effect of the presence of the lamb and suckling. Endocrinology. 147:3501-3509. Tobach E, Porto N. 2006. Note on social behavior of long-term captive female orangutans ( Pongo pygmaeus abellii ). Psychol Rep. 98:177-183. Torner L, Toschi N, Poh linger A, Landgraf R, Neumann ID. 2001. Anxiolytic and anti-stress effects of brain prolactin: Improved efficacy of antisense targeting of the prolactin receptor by molecu lar modeling. J Neurosci. 21:3207-3214. Touma C, Palme R. 2005. Measuring fecal gl ucocorticoid metabolites in mammals and birds: The importance of validatio n. Ann N Y Acad Sci. 1046:54-74. Tu MT, Lupien SJ, Walker CD. 2005. Measuring stress responses in postpartum mothers: Perspectives from studies in human and animal populations. Stress. 8: 19-34. Uvnas-Moberg K. 1998. Oxytocin may mediate the benefits of positive social interaction and emotions. Psychoneuroendo crinology. 23:819-835. Vallee M, Mayo W, Dellu F, Le Moal M, Simon H, Maccari S. 1997. Prenatal stress induces high anxiety and postnatal handling induces low anxiety in adult offspring: Correlation with stress-induced corticosterone secretion. J Neurosci. 17:2626-2636. van Belle S, Estrada A. 2008. Group size and composition influence male and female reproductive success in black howler monkeys ( Alouatta pigra ). Am J Primatol. 70:613-619. van Hooff JARAM, van Schaik CP. 1994. Male bonds: Affiliative relationships among nonhuman primate males. Behaviour. 130:309-337.
251 van Noordwijk MA, van Schaik CP. 1999. The e ffects of dominance rank and group size on female lifetime reproductive su ccess in wild long-tailed macaques, Macaca fascicularis Primates. 40:105-130. van Schaik CP. 1983. Why are diurnal prim ates living in groups? Behaviour. 87:120144. van Schaik CP. 1989. The ecology of social re lationships amongst female primates. In: Standen V, Foley RA, editors. Comparative socioecology: The behavioural ecology of humans and other mammals. Oxford: Blackwell Publishing. p. 195218. van Schaik CP, Aureli F. 2000. The natural hist ory of valuable relationships in primates. In: Aureli F, de Waal FBM, editors. Na tural conflict resolution. Berkeley: University of California Press. p. 307-333. van Schaik CP, van Hooff JARAM. 1983. On t he ultimate causes of primate social systems. Behaviour. 85:91-117. van Schaik CP, van Noordwijk MA. 1989. The special role of male Cebus monkeys in predation avoidance and its effect on group co mposition. Behav Ecol Sociobiol. 24:265-276. van Schaik CP, van Noordwijk MA, Deboer RJ Dentonkelaar I. 1983a. The effect of group size on time budgets and social behavior in wild long-tailed macaques ( Macaca fascicularis ). Behav Ecol Sociobiol. 13:173-181. van Schaik CP, van Noordwijk MA, van Braft T, Blakenstein MA. 1991. A pilot study of the social correlates of levels of urinar y cortisol, prolactin, and testosterone in wild long-tailed macaques ( Macaca fascicularis ). Primates. 32:345-356. van Schaik CP, van Noordwijk MA, Warsono B, Sutriono E. 1983b. Party size and early detection of predators in Sumatran fo rest primates. Primates. 24:211-221. Ventura R, Majolo B, Koyama NF, Hard ie S, Schino G. 2006. Reciprocation and interchange in wild Japanese macaques: Grooming, cofeeding and agonistic support. Am J Primatol. 68:1138-1149. Virgin CF, Sapolsky RM. 1997. Styles of male social behavior and their endocrine correlates among low-ranking ba boons. Am J Primatol. 42:25-29. Vogel ER. 2008. Tracking the forest people. Colby Mag. 96(3). Vogel ER, Janson CH. 2007. Predicting t he frequency of food related agonism in white-faced capuchin monkeys ( Cebus capucinus), using a novel focal tree method. Am J Primatol. 69:533-550.
252 Vogel ER, Munch SB, Janson CH. 2007. Under standing escalated aggression over food resources in white-faced capuchin monkeys. Anim Behav. 74:71-80. von Holst D. 1998. The concept of stress and its relevance for animal behavior. Adv Stud Behav. 27:1-131. Wadhwa PD. 2005. Psychoneuroendocrine processes in human pregnancy influence fetal development and health. Psyc honeuroendocrinology. 30:724-743. Walker SL, Smith RF, Jones DN, Routly JE Dobson H. 2008. Chronic stress, hormone profiles and estrus intensity in dairy cattle. Horm Behav. 53:493-501. Wasser SK, Hunt KE, Brown JL, Cooper K, Crockett CM, Bechert U, Millspaughn JJ, Larson S, Monfort SL. 2000. A generalized fe cal glucocorticoid assay for use in a diverse array of nondomestic mammalian and avian species. Gen Comp Endocrinol. 120:260-275. Watts DP. 1985. Relations between gr oup size and composition and feeding competition in mountain gorilla groups. Anim Behav. 33:72-85. Weingrill T, Gray DA, Barreu L, Henzi SP. 2004. Fecal cortisol levels in free-ranging chacma baboons: Relationship to dominance, reproductive state and environmental factors. Horm Behav. 45:259-269. Weinstock M. 2001. Alterations induced by gestational stress in brain morphology and behaviour of the offspring. Prog Neurobiol. 65:427-451. Whitten PL, Brockman DK, Stavisky RC. 1998. Recent advances in noninvasive techniques to monitor hormone-behavior in teractions. Yearb Phys Anthropol. 41:1-23. Wilkinson GS. 1986. Social groo ming in the common vampire bat, Desmodus rotundus Anim Behav. 34:1880-1889. Windle RJ, Wood S, Shanks N, Perks P, Conde GL, da Costa APC, Ingram CD, Lightman SL. 1997. Endocrine and behavioural responses to noise stress: Comparison of virgin and lactating fema le rates during non-disrupted maternal activity. J Neuroendocrinol. 9:407-414. Wingfield JC, Hunt R, Breuner C, Dunl ap K, Fowler GS, Freed L, Lepson J. 1997. Environmental stress, field endocrinol ogy and conservation biology. In: Clemmons R, Buchholz R, editors. Behavioral approaches to conservation in the wild. Cambridge: Cambridge Univ ersity Press. p. 95-131.
253 Wingfield JC, Maney DL, Breuner CW, Jac obs JD, Lynn S, Ramenofsky M, Richardson RD. 1998. Ecological bases of hormone-behavior interactions: The "emergency life-history stage". Am Zool. 38:191-206. Wittemyer G, Douglas-Hamilt on I, Getz WM. 2005. The socioecology of elephants: Analysis of the processes creating multit iered social structures. Anim Behav. 69: 1357-1371. Wittig RA, Crockford C, Lehmann J, Whitt en PL, Seyfarth RM, Cheney DL. 2008. Focused grooming networks and stress alleviation in wild female baboons. Horm Behav. 54:170-177. Wrangham RW. 1980. An ecological model of female-bonded primate groups. Behaviour. 75:262-300. Wright PC. 1994. The behavior and ecology of the owl monkey. In: Baer JF, Weller RE, Kakoma I, editors. Aotus : The owl monkey. San Diego: Academic Press. p. 97112. Zamma K. 2002. Grooming site preferenc es determined by lice infection among Japanese macaques in Arashi yama. Primates. 43:41-49. Zhang SY. 1995a. Activity and ranging patterns in relation to fruit utilization by brown capuchins ( Cebus apella ) in French Guiana. Int J Primatol. 16: 489-507. Zhang SY. 1995b. Sleeping habits of brown capuchin monkeys ( Cebus apella ) in French Guiana. Am J Primatol. 36: 327-335. Zhang SY, Wang L. 2000. Follo wing of brown capuchin m onkeys by white hawks in French Guiana. Condor. 102:198-201. Zhang SY, Wang LX. 1995. Fruit cons umption and seed dispersal of Ziziphus cinnamomum (Rhamnaceae) by two sympatric primates ( Cebus apella and Ateles paniscus ) in French Guiana. Biotropica. 27:397-401. Ziegler TE, Scheffler G, Snowdon CT. 1995. T he relationship of cortisol levels to social-environment and reproductive functioning in female cotton-top tamarins, Saguinus oedipus Horm Behav. 29:407-424. Ziegler TE, Sousa MBC. 2002. Parentdaughter relationships and social controls on fertility in female common marmosets, Callithrix jacchus. Horm Behav. 42:356367. Ziegler TE, Wittwer DJ. 2005. Fecal steroi d research in the field and laboratory: Improved methods for storage, transport, processing, and analysis. Am J Primatol. 67:159-174.
254 BIOGRAPHICAL SKETCH Erin Eliz abeth Ehmke graduat ed from The University of Florida in 1999 with a Bachelor of Science in wildlife ecology and conservation, a specialization in conservation and minors in zoology and envir onmental science. As director of Jungle Friends Primate Sanctuary in Gainesville, Florida from 1999 to 2001, Erins dream of working with primates became a reality and set in motion her future educational and career goals. Her experience studying wild primates began in Panama with The Institute for Tropical Ecology and Conser vation and continued with a year-long field assistantship studying brown capuchin monk eys in Raleighvallen, Suriname. Erin began her graduate education in Biol ogical Anthropology at The University of Florida in 2002 and earned her Master of Arts in 2004 with t he thesis study Social Interactions of Alpha, Natal and Immigrant Males with Juveniles among Brown Capuchins in Suriname. Erin then conti nued her education in the Depar tment of Anthropologys doctoral program at UF and returned to Suriname to conduct her dissertation research. Her Ph.D. dissertation, Stress and Affiliation among Wild Female Primates: Effects of Group Size, Risk, and Reproductive Condition in a Dynamic Forest Community, provides a direction of study that Erin plans to expand u pon, including the comparative effects of captive and natural environments on the behavioral endocrinology and socioecology of social mammals. The t eaching experience Erin acquired during her graduate career, through assistantships in the Departments of Zoology and Anthropology and as an instruct or with the Duke Talent Id entification Program (TIP), inspired a future in academia subsequent to the completion of her doctorate in 2010.