American .Journal of Primatology 3:277-281 (1983) BRIEF REPORT Outbreak of Severe Aggression in Captive Macaca mutatta AMY SAMUELS AND ROY V. HENRICKSON, California Primate Research Center, University of California, Davis An outbreak of severe aggression occurred among females in a rhesus macaque breeding group a t the California Primate Research Center four years after the group was established. During the breeding season in which this occurred, the incidence of injured females in other breeding groups a t the Primate Center was significantly higher than in the previous year. This breeding season was the first in which a large number of females reached sexual maturity. The group in which the most severe aggression occurred contained the largest number and proportion of maturing females. Evidence suggests that the simultaneous maturation of a large cohort of adolescent females may be associated with increased levels of aggression, and that this aggression may be intensified by certain aspects of captivity. Key words: aggression, domestic breeding, Macaca mulatta, rhesus macaques, sexual maturation INTRODUCTION The formation of a captive macaque group in which unfamiliar adults are confined together results in extensive aggressive behavior [Bernstein & Mason, 1963; Southwick, 1967; Bernstein, 19691. However, it is often assumed that once the initial aggression subsides, these groups are socially stable. Here we describe an outbreak of extreme aggression that occurred in a presumably stable group of rhesus macaques a t the California Primate Research Center (CPRC), and we discuss the factors that may have precipitated this aggression. METHODS CPRC has six field cage groups of rhesus macaques that share similar histories, group compositions, and management protocols [Hinshaw, 19811. Each breeding group is housed in a 0.2 hectare enclosure (30.5 m x 61 m), described in Neurater & Goodwin . The outbreak described in this paper occurred in Field Cage 7 (FC7), a group established in October 1976, with 47 wild-born monkeys, all of breeding age: 43 females of estimated age 3.5 to 8 years and four adult males. Although maternal kinship is known for the monkeys born in the group since formation, relatedness of the original members is not. Since 1976 several additions and removals of animals were made for management reasons: the replacement of a single remaining male with a group of adult and subadult males in 1977 [Sassenrath Received September 15, 1982; revision accepted July 3, 1983 Address reprint requests to Dr. Roy Henrickson, California Primate Research Center, University of California, Davis, CA 95616. 0 1983 Alan R. Liss, Inc. 278 Samuels and Henrickson & Goo, 19781; the experimental introduction of four captive-born juvenile females in 1977 [Sassenrath & Goo, 19781; and the subsequent removal of three of those females and five immature males in 1979. In late September 1980,20 immature males were removed from the group as part of long-term field cage management plans. All other demographic changes in the FC7 group were the result of births, deaths, and permanent removals of animals with chronic or severe health problems. At the time of the outbreak of aggression in August 1980, the group consisted of 120 monkeys: 51 infants and yearlings; 12 juvenile males; 20 adolescent females (2.0 to 3.5 years old); 33 adult females (> 3.5 years old); and four adult males. Behavioral observations were initiated as a response to the outbreak of aggression. Between August 25, 1980, and January 25, 1981, approximately 50 hr of observation of the FC7 group were conducted, during which details of all agonistic interactions were recorded ad libitum [Altmann, 19741. These data enabled us to identify the animals involved in mob aggression. In addition, these data were used to identify dominance relationships. For dyadic agonistic interactions in which one individual performed only aggressive behaviors (threaten, chase, bite, attack, or forcible displacement) and the partner performed only submissive behaviors (grimace, crouch, or flee), we scored the former as the dominant individual in the encounter and the latter as the subordinate one. Although we were able to produce a resultant dominance matrix for adult females, ordered to give a minimum number of only seven reversals out of 150 interactions, the existence of empty cells and few entries in some cells led us to treat dominance as a dichotomized variable-females were classified as either high-ranking or low-ranking. No reversals of rank occurred between females who were classified as high-ranking and those classified as lowranking. For eight females killed, or severely wounded and removed a t the start of the study, rank classification was assigned on the basis of a previous study (J. Simpson, unpublished data) for seven of them and on the basis of her adolescent daughter’s classification for the remaining one. Dominance ranks of the four adult males have been stable since 1977 [Sassenrath & Goo, 1978; D.G. Smith, personal communication]. RESULTS The Outbreak of Aggression in the FC7 Group Between August 22 and 27,22 monkeys were removed from the FC7 enclosure as a result of the outbreak of aggression. Ten animals were hospitalized and then returned to the group over a period of five weeks. Twelve were permanently removed from the group; seven of these died, one other was repeatedly injured seriously, and four infants were not returned to the group after the death or removal of their mothers. A disproportionate number (95%) of the monkeys who were killed or injured were female (x2=5.12, d f = l , p<0.05), and a disproportionate number (95%) of the female victims were adolescents or adults (x2=9.58, df= 1,p < 0.01). The only other injured animals were two infants who were inadvertently left in the enclosure overnight after the removal of their mothers. Victims often had maternal relatives who were also victims. The mother of each adolescent female who was injured or killed was herself a victim. Adult females who were victims were more likely than other adult females to have had at least one offspring who was also a victim (Fisher Exact p=0.003). Because family size may influence the probability that both a mother and one of her offspring would be victims, only the fourteen families that consisted of one adult female and three offspring were considered in this analysis; however, the finding was consistent for families of all sizes. Aggression in Khesus 279 Most victims were members of high-ranking matrilines (x2= 16.42, df= 1, p<O.OOl). All seven of the females who were killed and 12 of the 13 who were hospitalized were members of high-ranking families. Severe aggression occurred throughout the breeding season. This aggression tended to be sporadic and less intense than the initial outbreak of aggression in August. In 11(46%)of the 24 instances of mob aggression directed toward a female, the attacked female had been a victim of the August outbreak or was the daughter of an August victim. In two thirds of mob attacks, the attacked female was an adolescent. Victims were observed to sustain wounds in 2 of the 24 mob attacks. Both times, the wounded monkey was a female who had been a n August victim. In 88% of mob attacks, all aggressors in the mob were adolescent or adult females. Seventy-one percent of attacks included at least one adolescent female aggressor, and 96% included a t least one high-ranking female aggressor. Of a total of 74 aggressors identified in the mob attacks, 76% were high-ranking females, 19% were low-ranking females, and 5% were males. A male was a n aggressor in only three attacks; in the other 12 instances of male involvement, the adult male “protected” the female victim [Bernstein, 19641. No aggression by a male directed toward a female was observed to result in injury. In summary, aggression was observed primarily among high-ranking females, and both attackers and victims were often adolescent females. Comparisons Between the FC7 Group and the Other Field Cage Groups Because all of the field cage groups were formed in 1976 or later, the first cohort of infants was born in 1977. Consequently, the 1980 breeding season was an unusual one for all of the field cage groups. It was the first in which many females reached sexual maturity. The maturation of this cohort of females had a dramatic effect on relative proportions of females of different age classes in the field cage population. In 1979, 13% (34/201) of all potentially sexually cycling females ( > 2 years) were adolescents, whereas in 1980 nearly one third (731231)were so. Although initial compositions of the six field cage groups were similar, the FC7 group had fewer health problems and was more reproductively successful than the other groups. As a result, by August 1980 the FC7 group was 43% larger than the next largest field cage group, this difference due to the large number of females in all ages classes. In particular, a t the onset of the 1980 breeding season, the FC7 group had 26% more females of potentially cycling age than did any other group. The other five groups exhibited a similar but less dramatic increase in the incidence of wounds. Between August and January the rate of injury (= number of animals hospitalized for wounds I total number of monkeys) was significantly higher in 1980-81 than in 1979-80 (Test of proportions, p =0.0003). A disproportionate number of the 44 monkeys from the other groups who were hospitalized for injuries during the study period were adolescent females (32%) and adult females (50%) (x2=10.86, d f = l , p<O.OOl). As had been the case in the FC7 group, mothers of injured adolescents were often victims themselves. Of all the adolescent females who had mothers still living, significantly more who were victims had mothers who were also victims than would be expected by chance (N=42 juveniles living with their mothers; x2 =4.33, df= 1, p < 0.05). The number of monkeys hospitalized per group was not a function of group size (Spearman rank correlation, p > 0.15), and the smallest and largest groups had the highest rates of injury. Moreover, all other groups had higher proportions of injured monkeys in 1980 than did the FC7 group in 1979, even though none of those groups was as large in 1980 as the FC7 group was in 1979. In contrast with the absence of a relationship between injury rate and group size, the relationship between injury 280 Samuels and Henrickson rate and number of adolescent females approached statistical significance (r, p=0.07). = 0.70, DISCUSSION That the victims in the present study were female kin and the aggressors were females suggest that the aggression occurred between matrilineal groups. This is not surprising in view of what is known of the pervasive influence of kinship on the behavior of female macaques [Kurland, 19771. Within the CPRC population, the FC7 group was larger than any other group a t the time of the outbreak of aggression. This may explain, in part, why the increase and pattern of interlineage aggression, common to all the groups, was expressed most dramatically in this group. However, macaque groups that are larger and living a t higher densities than the FC7 group have been reported, and comparable outbreaks of aggression have not been reported for them (R.Wolf, personal communication). Particularly relevant to the present results are those for groups that experienced changes in “social density” [Nagel & Kummer, 19741 or group composition over a period of several years. In a captive group of Macaca fuscata that had recruitment only through births, aggression among adults decreased as group size increased from 107 to 192 individuals over a five-year period (Eaton et al, 19811. Thus, stress that is likely to result in severe aggression may be a function of demographic variables other than, or in addition to, size or density. The maturation of a large cohort of females after a hiatus in maturations has been associated with increased aggression in a group of wild Papio cynocephalus [Altmann & Altmann, 1979; Walters, 19801. In that group, the previously stable female dominance hierarchy was disrupted, and a n outbreak of unusually severe aggression among high-ranking females was observed when the maturing females began challenging adult females who ranked below their mothers (J. Walters, personal communication). Increased levels of aggression as a result of female maturation may be exacerbated in captivity where monkey groups are often established with unfamiliar sexually mature animals, and initial dominance relations are established in the absence of kinship ties. Under such conditions there is no recruitment, particularly of related animals, into the adult class until the first birth cohort attains reproductive age. This contrasts with newly formed groups that form by fissioning along family lines in which all ageisex classes are represented [Missakian, 1973; Chepko-Sade & Sade, 1979). Moreover, as a result of enhanced fertility and survival in captivity [Smith, 19821, the size of maturing cohorts will usually be larger in captivity than those produced by the same number of females in free-ranging populations. Behavioral patterns that evolved under demographic conditions that change gradually, in wild populations, may have a different impact when such changes occur rapidly, in captivity. The present analyses suggest that the simultaneous maturation of a large cohort of females can be associated with severe aggression. Perhaps more important, these results emphasize the need for exchange of information between behavioral researchers and those managing breeding colonies of monkeys [Bernstein & Gordon, 19771. Ongoing behavioral monitoring schemes and greater attention to demographic factors other than total size and density will add to the productivity and quality of conditions in captive populations, as well as increase basic understanding of factors that facilitate and limit severe aggression. ACKNOWLEDGMENTS We thank L. Berenstain, G. Eaton, K. Hinshaw, M. Leighton, M. Pereira, M. Rowe, J. Simpson, J. Walters, Mmu 11, and members of the Allee Lab and CPRC Aggression in Rhesus 281 staff for logistical support, discussion of the ideas presented, and comments on drafts of this manuscript. In particular, we thank J. Altmann and J. Silk for invaluable assistance in putting together the pieces of the puzzle. This research was supported by U.S. Public Health Service Grant RR00169 from the National Institutes of Health. REFERENCES Altmann, J. Observational study of behavior: sampling methods. BEHAVIOUR 49: 227267,1974. Altmann, S.A.; Altmann, J. Demographic constraints on behavior and social organization, pp. 47-64 in PRIMATE ECOLOGY AND HUMAN ORIGINS: ECOLOGICAL INFLUENCES ON SOCIAL ORGANIZATION. I.S. Bernstein; E.O. Smith, eds. New York, Garland STPM Press, 1979. Bernstein, I.S. Role of t h e dominant male rhesus i n response to external challenges to the group. JOURNAL O F COMPARA?IVE PHYSIOLOGY AND PSYCHOLOGY 57: 404-406, 1964. Bernstein, I.S. Introductory techniques in the formation of pigtail monkey troops. FOLIA PRIMATOLOGICA 10: 1-19, 1969. Bernstein, 1,s.; Gordon, T.P. Behavioral research in breeding colonies of old world monkeys. LABORATORY ANIMAL SCIENCE 27:532-540, 1977. Bernstein, I.S.; Mason, W.A. Group formation by rhesus monkeys. ANIMAL BEHAVIOR l i : 28-31,1963, " CheDko-Sade.B.D.: Sade. D.S. Patterns of &ups splitting within matrilineal kinship &oups. BEHAVIORAL ECOLOGY AND SOCIOBIOLOGY 5 : 67-86.1979. Eaton, G.G.; Modahl, K.B.; johnson, D.F. Aggressive behavior in a confined troop of Japanese macaques: effects of density, season and gender. AGGRESSIVE BEHAVIOR 7:145-164, 1981. Hinshaw. K.C. Vital statistics for multimale breeding groups of rhesus monkeys (Macaca rnulatta) housed in outdoor field cages at the California Primate Research Center, 1976- 1980. MASTERS THESIS, University of California, Davis, 1981. Kurland, J.A. Kin Selection in the Japanese Monkey. CONTRIBUTIONS TO PRIMATOLOGY 12, Basel, S. Karger, 1977. Missakian, E.A. The timing of fission among free-ranging rhesus monkeys. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 38: 621-624,1973. Nagel, U.; Kummer, H. Variation in cercopithecoid aggressive behavior, pp. 159-184 in PRIMATE AGGRESSION, TERRITORIALITY, AND XENOPHOBIA. R.L. Holloway, ed. New York, Academic Press, 1974. Neurater, L.J.; Goodwin, W.J. The development and management of macaque breeding proggams, pp. 60-75 in BREEDING PRIMATES, W.I.B. Beveridge, ed. Basel, Karger, 1972. Sassenrath, E.N.; Goo, G.P. Augmentation of established domestic breeding groups of rhesus macaques. PROCEEDINGS OF 29TH ANNUAL SESSION OF AMERICAN ASSOCIATION OF LABORATORY ANIMAL SCIENCE, New York, 1978. Smith, D.G. A comparison of the demographic structure and growth of free-ranging and captive groups of rhesus monkeys (Macaca mulatta). PRIMATES 23: 24-30, 1982. Southwick, C.H. An experimental study of intragroup agonistic behavior in rhesus monkeys (Macaca mulatta). BEHAVIOUR 28: 182-209, 1967. Walters, J . Interventions and the development of dominance relationships in female baboons. FOLIA PRIMATOLOGICA 34: 6189, 1980.