American Journal of Primatology 23:171-183 (1991) Homosexual Incest Avoidance Among Females in Captive Japanese Macaques BERNARD CHAPAIS AND CHRISTIANE MIGNAULT Dkpartement d'Anthropologie, Uniuersite de Montreal, Montreal, Canada This paper reports a systematic pattern of homosexual incest avoidance among females in a captive group of Japanese macaques (Macaca fuscata) culled from the Arashiyama-West troop known for its high rates of female homosexuality. The study group included three matrilines and two generations. Between eight and 11 females were sexually active over four consecutive mating seasons, and all engaged in both heterosexual and homosexual activity. While all females performed homosexual acts with almost all possible non-kin partners, they systematically avoided homosexual interactions with their mother, daughters, and sisters. This pattern could not be explained either in terms of kin not being simultaneously in estrus, kin avoiding affiliative interactions in general, or non-kin utilizing the tension-reducing effect of estrus to affiliate exclusively with each other. In contrast to homosexual females, heterosexual pairs of relatives (brother-sister, mother-son) were sometimes incestuous. Assuming that female homosexuality expresses the reproductive strategy of females unconstrained by male influence, the present results point to the strong tendency of females to avoid incest and suggest that males are primarily responsible for the reported exceptions to incest avoidance. Key words: female homosexuality, sexual behavior, Mucucu fuscata INTRODUCTION Incest avoidance between parents and offspring, and between siblings, is a relatively well-documented phenomenon in nonhuman primates. The fact that in most primate species the members of one sex, or of both, leave their natal group around puberty [Pusey & Packer, 19871 accounts for an important part of that phenomenon. But even in situations in which inbreeding would be possible, for example in the case of mature males lingering in their natal troop [Koford, 1963; Tilford, 1982; Chapais, 1983al or in captive groups (dispersal prevented), intrafamilial incest appears to be generally avoided. Mother-son mating either has not been observed [Tokuda, 1961-1962; Baxter & Fedigan, 1979; Wolfe, 1979; Pusey, 1980; Paul & Kuester, 1985; Smuts, 1985; Taylor & Sussman, 19851 or has been reported to take place at low frequencies [Loy, 1971; Sade, 1968,1972; Itani, 1972; Received for publication February 14,1990; revision accepted October 15, 1990. Address reprint requests to Bernard Chapais, Departement d'Anthropologie, Universite de Montreal, C.P. 6128, Succ. A, Montrbal, Canada H3C 357. 0 1991 Wiley-Liss, Inc. 172 / Chapais and Mignault Enomoto, 1978; Murray & Smith, 19831. Similarly, brother-sister mating either has not been reported [Baxter & Fedigan, 1979; Chapais, 198313; Paul & Kuester, 1985; Taylor & Sussman, 19851 or has been observed at low frequencies [Sade, 1972; Wolfe, 1979; Pusey, 1980; Goodall, 19861. Moreover, when mother-son or brother-sister matings do occur, they usually involve immature or adolescent males who may not be fertile [Loy, 1971; Sade, 1972; Missakian, 1973; Savage & Malick, 19771. Father-daughter incest appears infrequent as well in monogamous species [e.g., Abbott, 19841, one-male groups [Kummer, 1968; Dunbar & Dunbar, 19751, and multi-male groups [Packer, 1979; Pusey, 1980; Scott, 1984; Smuts, 1985; Goodall, 19861. Incest avoidance, therefore, appears to be a general characteristic of heterosexual activity in nonhuman primates. In the present paper, we are concerned with the extension of incest avoidance into the realm of homosexual activity. Female homosexuality has been observed in many primate species [Akers & Conaway, 19791. In some species, homosexual females exhibit behaviors typical of heterosexual activity. They form consortships lasting between hours and days during which they mount, groom, and huddle. It is noteworthy that homosexual females are not exclusively so; they also engage in heterosexual activity. In Japanese macaques, the proportion of females engaging in homosexual activity over a given mating season varied between troops and between years. For example, in the ArashiyamaWest troop (Texas), that proportion varied between 78% [Wolfe, 19791 and 51% [Gouzoules and Goy, 19831. In the Arashiyama B troop (Japan),it varied between 47% [Takahata, 19821and 27% [Wolfe, 19861,and in the Oregon troop, it was lower than 12% over four mating seasons [Eaton, 19781. Wolfe [1984,1986] has proposed that high rates of female homosexuality express a quest for sexual novelty and are ultimately related to alterations in the sex ratio. Some data suggest that incest avoidance extends to homosexual activity. For example, in Japanese macaques, Wolfe [19791, Baxter & Fedigan [19791, and Gouzoules & Goy [ 19831 reported no homosexual consortships between mother and daughters and none between sisters. On theoretical grounds, the issue of the relation between incest avoidance and homosexual activity may contribute t o a better understanding of the proximate and ultimate causes of the phenomenon of incest avoidance in general. In heterosexual activity, mate selection reflects a compromise between the reproductive strategies of each sex (i.e., between female selectiveness and male promiscuity, according to sexual selection theory). On the other hand, in the case of homosexuality, Symons [19791 has suggested that mate selection reflects primarily the mating strategy of'the sex involved. This author has interpreted several contrasting characteristics of human male and female homosexuality in that perspective (number of partners, faithfulness, etc.). According to this hypothesis, female homosexuality would express female choice in the absence of any influence by males. Under this assumption, the analysis of the identity of homosexual female partners might allow one to isolate the attitude of females toward incest. Similarly, male homosexuality might make it possible to study male choice of sexual partners in the absence of female influence. Parental investment theory states that in mammals females should have a stronger interest than do males in the avoidance of incest because they generally invest more than do the latter in each offspring and have more to lose from inbreeding depression [Trivers, 1972; Clutton-Brock & Harvey, 19761. Combining this principle with Symons's [ 19791 hypothesis, we would expect female homosexuality to reflect a degree of incest avoidance more pronounced than the degree characterizing homosexual males or heterosexual partners. Homosexual Incest Avoidance in Macaques I 173 TABLE I. Sexually Mature Individuals, Number of Observation Days, and Mean Number of Observation Hours Per Day for Each of the Four Mating Seasons Data collection Sexually mature individuals Mating season No. Hours/ K A A1 A2 A3 A4 B B1 B2 B3 B4 B5 C C1 C3 C4 davs dav 1985-86 1986-87 1987-88 1988-89 m m m m f f f f f f f f f f f f f f f f f m f f f f f f f f f f f f m f f f f f f f f f f f f m 1 57 86 182 1 5 2.5 2 5.5 3.5 m: male: f: female In this paper, we present data on female homosexuality in a captive subgroup of Japanese macaques culled from the Arashiyama-West troop known for its high rate of female homoxexuality (see above). The rate of female homosexuality was particularly high in the study group, a situation which provided a quantitatively solid basis for the present analysis. METHODS Subjects A group of 15 Japanese macaques from the Arashiyama-West troop (translocated to Texas; see Clark & Mano ) was established at the Universite de Montreal in the summer of 1984. This group consisted of an adult male (K) and three matrilines with similar age-sex compositions. In each matriline, the mother (named A, B, or C) was born in 1971 or 1973 and had three immature daughters, born in 1981, 1982, and 1983, and one son born in 1984 (the 1982 daughter of C could not be obtained). One female died (Al, in April 1988), and 16 individuals have been born since then. The A and C families were members of the same genealogy (Betta), but were distantly related through the great-grandmother of A and C. The B matriline belonged to another genealogy (Rotte), and the adult male, to a third one (Kujiro). The A-B-C notation refers to the dominance rank of matrilines (see below). The present analysis focuses on the four mating seasons extending between the autumn of 1985 and the winter of 1989. Mating seasons ranged between September and March, with some yearly variation relating to the number of pregnant females. Sexually mature individuals during each of these seasons are listed in Table I. Females had their first estrus a t 3.5 years so that between eight and 11 females were sexually active depending on the mating season. Note that female A1 died before the onset of the fourth mating season. The sons of the three matriarchs (A4, B4, and C4) were all sexually active (series mounting) at 3.5 years of age (third mating season), but were sexually mature (i.e., ejaculated) at 4.5 years of age (except for B4 who ejaculated a few times at 3.5 years). Housing The monkeys occupied two indoor rooms (7.2 x 4.4 x 3 m; 4.7 x 3.3 x 3 m) and an outdoor pen (10 x 3 x 5 m) throughout the year. From June 1987 on, they had access to a third indoor room (6 x 5 x 3 m). All rooms and the outdoor pen were equipped with cables and suspended boards. The animals were fed daily with monkey chow and various types of seeds distributed in the deep woodchip litter of the indoor rooms. Fruits and vegetables were distributed every day, and water was available ad libitum. 174 / Chapais and Mignault Behaviors Homosexual females mounted each other, in general reciprocally, using different positions: double-foot clasp, sitting on their partner’s back, or lying on it. They often performed pelvic thrusts or rubbed their clitoris on their partner, often looking at and pulling each other’s fur. In order to prompt a mount by their partner, they could slap the ground, emit high-pitched vocalizations, present their hindquarters, or run away and come back presenting. Between mounts, they usually stayed in proximity, groomed each other, and huddled. Such associations extended over several days and weeks and were usually intense (as evidenced, for example, by the number of mounts, see below). Consortships were terminated with a change of partner or with the end of estrus. Females engaging in homosexual activity had all the characteristics of estrous females: reddening of the face and perineum, perineal swellings in the case of 3 year olds, high rates of sexual vocalizations, all behavioral components typical of heterosexual activity, higher activity levels, etc. Thus, the form and temporal patterning of homosexual interactions had much in common with heterosexual activity (for further details, see Wolfe [ 1979, 19841). An homosexual dyad was defined as one in which the females mounted each other repeatedly (regardless of the type of mounts). Since this paper is concerned with the distribution of homosexual activity among females, not with homosexual behavior per se, only the frequencies of mounts were analyzed. This criterion appeared to be the most significant because females who mounted each other always performed the other behaviors mentioned above. On the basis of the frequencies of mounts, we calculated the number of days in consort for each female pair for each season and the number of homosexual partners per female. Dominance relations were determined on the basis of the direction of submissive behaviors (supplantations, fear grimaces, and flights). Data on the direction of inter-familial dominance relationships in the whole group are given elsewhere [Chapais, 1988; Chapais et al., in press]. Over the four mating seasons, the members of the B matriline submitted to the members of the A matriline and the members of the C matriline submitted to both the A and B matrilines. Data Collection The monkeys were observed from a central observation booth. The observability was excellent. Data were coded on paper in a temporal sequence marked by l-minute intervals. Table I gives the number of observation days and observation hours per day for each of the mating seasons. Data collection extended between 08:30 hours and 16:OO hours on weekdays and between 1O:OO hours and 14:OO hours during weekends. During the first three mating seasons, focal sampling on each of the sexually mature females was carried out. Sexual activity taking place during focal samples, but not involving the focal individual, was also noted systematically. Because observability was excellent and sexual interactions are typically long-lasting, such concurrent, behavior-specific sampling yielded information on all the dyads engaging in sexual activity as well as a very good sample of that activity for each dyad. During the fourth mating season, due to the requirements of another research project, sexual activity was noted on an ad libitum basis in the course of focal samples on immature individuals. Again, due to good observability and the long duration of sexual interactions, all dyads engaging in sexual activity could be identified. However, the sample of behaviors for each dyad was considerably reduced. Accordingly, the present data from the fourth mating season are limited to Homosexual Incest Avoidance in Macaques I 175 TABLE 11. Percentage of Observation Days on Which Each Female Was Homosexual (Regardless of Whether She Was Also Heterosexual on That Day) and Heterosexual (Regardless of Whether She Was Also Homosexual on That Day) Mating seasons 1985-1986 1986-1987 1987-1988 1988-1 989 Homosexual Heterosexual Homosexual Heterosexual Homosexual Heterosexual Homosexual Heterosexual A A1 A2 A3 31.5 64.9 49.1 24.6 1.2 1.8 a 8.8 43.9 38.6 17.5 3.5 3.5 b b 18.1 6.0 2.2 23.1 17.0 1.1 3.8 16.5 20.0 B 31.9 17.0 45.6 40.1 25.8 3.8 26.4 43.4 a b 14.0 4.7 0.0 7.0 4.7 9.3 1.2 4.7 13.0 b 41.9 64.0 94.2 58.1 1.2 41.9 79.1 34.9 32.2 24.3 6.1 12.2 7.8 9.6 23.5 9.6 12.2 15.7 12.2 12.2 1.7 6.1 6.1 13.9 4.3 14.8 13.9 3.5 Females B1 B2 B3 B5 C c1 c3 b 0.0 61.4 b b 3.5 1.8 b b 27.9 37.2 50.0 b 8.1 1.2 1.2 b b 0.0 22.0 53.3 8.8 4.4 13.2 aAlhad died. 'The female was sexually immature. the identity of sexual partners (i.e., data on the frequencies of mounts come from the first three mating seasons only). A total of six different observers collected data on sexual activity over the four mating seasons. Four of them did so over more than one mating season. All were very familiar with the monkeys. Measures of inter-observer agreement were taken. In this project, only the mounts were analyzed and the percentages of agreement among observers for that category of behaviors were 100%. RESULTS Kinship and the Distribution of Homosexual Activity Among Females All females engaged in both heterosexual and homosexual activity over the four mating seasons (Table 11).This indicates that homosexuality was not confined to any mating season and that it did not constitute an exclusive type of sexual activity for any female, matriline, or age group. Table I11 presents the distribution of mounts among sexually mature females over the first three mating seasons. Although all females mounted other females, none of the kin dyads (mother-daughter, sister-sister) were homosexual (N = 30 cells), whereas 82.5% of the non-kin dyads (N = 80 cells) engaged in homosexuality. Table IV gives the total number of days during which each pair of females engaged in homosexual activity. This table includes the fourth mating season (see Methods) and, thus, one additional female who had her first estrus (B5). Again, none of the 19 kin dyads were homosexual, whereas 78.7% of the 47 non-kin dyads were homosexual (x2 = 34.0, df = 1,P < 0.001). Females mounted an average of six females (range four to seven) out of a maximum of seven non-kin partners for an A or B female or eight non-kin partners for a C female. Mounts were bidirectional within dyads (distributed on both sides of the diagonal in Table 111). Of the 10,035 mounts recorded, 57.4% were performed by the dominant female of the pair, and 42.6%, by the subordinate female. There were 21 dyads in which the dominant female mounted the subordinate female more often 176 I Chapais and Mignault TABLE 111. Distribution of Homosexual Mounts Among Sexually Mature Females Over the First Three Mating Seasons Mountees Mounters A A1 A2 A3 \ A A1 A2 A3 B B1 B2 B3 C 44 153 127 201 9 111 199 c1 c3 166 164 412 183 41 202 125 255 149 303 238 31 39 138 211 404 1121 195 213 B B1 B2 B3 C C1 C3 100 107 165 159 63 167 126 145 95 113 40 123 51 26 82 132 270 320 72 79 425 166 19 401 99 120 56 22 66 100 57 TABLE IV. Total Number of Days on Which Each Pair of Females Engaged in Homosexual Activity Over the Four Mating Seasons A 21 A2 A3 B B1 B2 B3 B5 C Cl c3 A1 \ A2 A3 B B1 B2 B3 22 7 25 19 25 43 41 40 45 19 37 52 11 29 9 B5 3 C C1 C3 10 9 6 4 7 17 42 39 18 19 85 28 4 8 3 17 26 16 11 29 \ ' 1 than the reverse and 13 dyads in which the subordinate female mounted more often. This difference is not significant (Wilcoxon test, N of pairs = 34, Z = 1.71, two-tailed P > 0.05). In summary, females engaged in homosexuality with almost all possible extrafamilial partners, but never with their mother, daughters, and sisters. The direction of mounting was not related to the relative rank of the two partners. Alternative Explanations to Incest Avoidance Before concluding that the absence of homosexuality among matriline members reflects a pattern of incest avoidance, we must eliminate at least three alternative explanations. First, this absence might simply be an artifact of the timing of estrous periods. It might result from the possibility that any two members of the same matriline had never or rarely been in estrus simultaneously, i.e., the estrous periods of non-kin females would have overlapped, while those of kin would not have. Table V shows that this was not the case. The members of the same matriline often had overlapping estrous periods. The mean number of days during which any female pair were simultaneously in estrus was 48.9 for non-kin females and 40.8 Homosexual Incest Avoidance in Macaques / 177 TABLE V. Number of Days During Which Each Pair of Females Were Simultaneously in Estrus During the Four Mating Seasons A A1 A2 A3 B B1 B2 B3 B5 C C1 C3 52 108 26 41 34 45 69 38 82 119 84 43 109 94 - 10 9 15 27 17 61 100 69 73 59 123 103 17 42 1 5 2 4 9 18 24 40 69 ~ A A1 A2 A3 B B1 B2 B3 B5 C c1 c3 32 27 82 91 29 13 64 \ Al-B5 dyad: at the time B5 had her first estrus (fourth mating season), A1 had died. for females of the same matriline. This difference is not significant (Mann-Whitney test, Z = 0.945, P > 0.05). Thus, members of the same matriline did not engage in homosexual activity with one another despite the fact that they were often simultaneously in estrus. Second,the absence of intra-matrilineal homosexuality might be the extension of a more global pattern of low affiliation rates among kin in this particular group. It could be that the members of the same matriline affiliated very little, either during the mating or during the non-mating season. Table VI presents the distribution of grooming among females during one mating season, and Table VII, during one non-mating season. In marked contrast to the absence of mounts among matriline members (Table 111),grooming was frequent among kin both during the mating and the non-mating seasons. In fact, grooming was much more frequent among kin than among non-kin over both seasons (Mann-Whitney test; mating season: Z = 4.29, P < 0.001;non-mating season: Z = 4.53, P < 0.001). These data indicate that the absence of homosexuality among female kin was not simply one manifestation of a more general tendency by kin to avoid affiliation. Third, the absence of female homosexuality within matrilines might reflect not so much the fact that kin actively avoided each other, but that the context of estrus allowed females to fully express their attractiveness to non-kin females. Assuming 1) that non-kin females are attracted to each other in general, 2) that tension is greater among non-kin females than among kin, and 3) that such tension is more easily overcome (i.e., females are less inhibited) in a context of sexual receptiveness, homosexuality would be expected to be more frequent among nonkin. In other words, females might utilize the context of estrus to affiliate with non-kin. A number of considerations, however, suggest that this hypothesis does not fully account for the absence of intra-matrilineal homosexuality in the present group. Assuming that tension is indeed lower among female kin (and knowing that kin are attracted to each other in general), one would not expect them to avoid specifically sexual interactions, unless some form of inhibition is present. On the other hand, the available evidence indicates that some degree of tension is likely to be present in all social relationships, kin and non-kin. For example, intramatrilineal competition for dominance among females is often pronounced in mon- 178 / Chapais and Mignault TABLE VI. Distribution of Grooming Among Females During the 1986-1987 Mating Season G roome e s Gr o o me r s A A1 A A1 A2 A3 B B1 B2 B3 C c1 c3 2 13 1 30 49 52 1.7 12 1.6 A2 A3 31 40 31 94 12 52 B B1 B2 B3 C 16 10 36 17 49 1 10 54 68 73 10 11 29 37 12 46 12 14 80 C1 C3 18 38 Based on 3,000 on-the-minute scan samples (50 hours) collected over 20 days between October 1 and December 16, 1986. Figures are frequencies of scan samples with grooming. TABLE VII. Distribution of Grooming Among Females During the Non-mating Season of 1985 G roome e s Gr o o me r s A A1 A2 A3 B B1 B2 B3 C c1 c3 A 23 A1 31 94 74 A2 A3 86 36 13 16 38 B El B2 B3 13 13 17 3 8 13 78 C C1 C3 38 18 16 89 51 24 6 3 58 4 16 11 9 2 8 6 20 8 37 17 31 4 23 10 33 5 17 22 32 13 68 Based on 2,400 on-the-minute scan samples taken over 20 days between J une 3 and July 5, 1985. Figures are frequencies of scan samples with grooming. keys [e.g., Lee, 1987; Datta, 19881, and the present group was no exception to that rule (personal observation). Thus, kin, as well as non-kin, would be expected to benefit from the tension-reducing effect of estrus t o affiliate with each other. This is hard to reconcile with the observed total absence of sexual interactions among female kin. Also, according to the above hypothesis, rates of affiliation among non-kin females would be expected to be significantly higher during estrous periods compared to periods outside of these. However, the incest avoidance hypothesis makes the same prediction and such a comparison, therefore, would not constitute a valid test of the present hypothesis. Lastly, still according to the present hypothesis, most of the tension between any two non-kin females would be expected to be felt by the subordinate female of the pair. On this basis, one would predict that the lower-ranking female would be considerably more active than would her dominant Homosexual Incest Avoidance in Macaques J 179 TABLE VIII. Heterosexual Activity of the Male Members of the Three Matrilines Mating Males A4 B4 c4 season 1987-88 1988-89 1987-88 1988-89 1987-88 1988-89 Females A A1 A2 A3 10 5" 10 5 - 6" 1 4" 2 3 B 1 1 1 8" 2" 33 7 B1 B2 2 3 2 1 1 B3 2 4" 4 1 1 15 6" B5 C - - - 3" 4 6" 5 1 C1 C3 1 1 5 5" 1 9" 12" 2 2 Figures are the number of days on which the male series mounted the female at least once. B5 was not sexually active during the 1987-1988 mating season, and A1 died prior to the 1988-1989 mating season. "The male was seen ejaculating at least once. partner in initiating and maintaining homosexual interactions. This prediction, however, is not supported by the present data. The frequencies with which female partners initiated homosexual interactions by approaching each other (within an arm's reach) were calculated for the best-documented mating season (the second one). Of all approaches recorded (N = 2,8831, 52.2% were performed by the dominant female of the pair (N = 23 pairs) and there was no significant difference between the number of approaches by the dominant and the subordinate females (Wilcoxon test; N = 21, T = 94, P > 0.05). Overall, these arguments render unlikely the hypothesis that the tension-reducing effect of estrus fully accounts for the absence of homosexuality among members of the same matriline. Kinship and the Distribution of Heterosexual Activity Table VIII presents data on the heterosexual activity of the male members of the three matrilines when they became sexually active at 3.5 years and sexually mature at 4.5 years. Each male series mounted many females, including his sisters (and mother in the case of C4). Non-incestuous dyads, however, were represented significantly more frequently than were incestous dyads: 45.5%of the 22 possible male-female kin dyads engaged in sexual activity, whereas 72.7% of the 44 non-kin dyads did so (x2 = 4.71, df = 1,P < 0.05). Non-kin dyads also mated on a greater number of days than expected on the basis of the proportion of kin and non-kin dyads (x2 = 31.1, df = 1,P < 0.001). DISCUSSION The high rates of female homosexuality observed in the present group (Table 11) probably reflected the combination of two factors: 1) the history of female homosexuality characterizing the Arashiyama-West troop [Wolfe, 1979, 1984, 19861, of which the study group was a subset; and 2) the fact that the study group included only one adult male, who was aged only 6 years a t the onset of the first mating season. This atypical situation provided a quantitatively solid basis for the analysis of the effect of kinship on partner selection among homosexual females. Although all females had many non-kin homosexual partners, they systematically avoided intra-matrilineal partners. Such selectivity could not be fully accounted for in terms of kin not being simultaneously in estrus, kin avoiding affiliative interactions in general, or non-kin utilizing the tension-reducing effect of estrus to affiliate exclusively with each other. We conclude, therefore, that the selectivity of females expressed a pattern of homosexual incest avoidance. This interpretation does not exclude the possibility that homosexual females derive 180 / Chapais and Mignault non-sexual benefits from affiliating with non-kin partners. For example, homosexual consortships sometimes led to temporary rank changes, whereby a low-ranking female (e.g., C l ) outranked the females subordinate to her high-ranking female partner (e.g., A). Interestingly, homosexual females always avoided incest, but heterosexual pairs were sometimes incestuous (brother-sister, mother-son; Table VIII). Although incestuous matings occurred significantly less often than did non-incestuous ones (in agreement with previous reports of heterosexual incest avoidance), the mere occurrence of heterosexual incest contrasts strongly with the total absence of homosexual incest among females. Thus, when females were the only potential partners involved, they always avoided incest, whereas when the two sexes were involved, exceptions did occur. Two explanations for this contrast are possible. According to the first of these, which is based on the theory of sexual selection and parental investment [Trivers, 19721, the above data would support the idea that males should be generally less selective than are females and, thus, less reluctant to mate with their kin. Accordingly, males would sometimes initiate sexual interactions with their kin and be primarily responsible for the observed cases of incest. In contrast, the pattern of complete incest avoidance found among homosexual females would reflect the fact that females are more selective and that they typically do not initiate incestuous interactions. Some evidence supports this hypothesis. Limited data on the role of the two sexes in the initiation of incestuous matings were available for the third mating season, when the three males were sexually active (no quantitative data were available for the fourth mating season; see Data Collection). Because males A4 and B4 were seldom incestuous during the third mating season (Table VIII), only the data for C4 could be analyzed. There were a total of 287 approaches between C4 and females unrelated to him on the days he series-mounted the latter; these females initiated 25.6% of the approaches. In contrast, C4’s three female kin never approached him on the days he series-mounted them, whereas he approached them 99 times on these days. Other authors have noted that the responsibility of males in initiating heterosexual incest is greater than that of females. For example, male chimpanzees sometimes try to copulate with their sisters or mother, but the latter avoid them [Goodall, 19861. Smuts  has reported that young female baboons are less attracted by their mother’s male friend (potential father) than by other males and that they respond less to their sexual advances. These findings contrast with the active role of females in the initiation of normal heterosexual matings [e.g., Smuts, 19871. The foregoing hypothesis refers to Symons’s [19791 idea that homosexual behavior expresses the reproductive strategy of the sex involved (in this case, female selectiveness), while heterosexual activity reflects a compromise between the male and female interests. It should be noted that this explanation implies that females blindly apply a phylogenetically built program of heterosexual incest avoidance in the context of homosexual interactions. More specifically, their behavior may be guided by the proximate mechanism that supposedly underlies incest avoidancethe Westermarck [ 18911 effect: Early and prolonged familiarity between individuals would foster sexual repulsion among them. An alternative explanation for the total absence of homosexual incest among females makes no reference to the differential reproductive strategies of males and females. It relates to the Westermarck effect only. If, in the course of ontogeny, females affiliate with their female kin more extensively than they do with their male kin, the Westermarck effect would operate more intensely among female kin Homosexual Incest Avoidance in Macaques / 181 than it would between male and female kin. Some data lend support t o this hypothesis. In savanna baboons and rhesus macaques, immature females affiliate more with some of their female kin than with some of their male kin [Ehardt & Bernstein, 1987; Pereira, 19881. These data are limited, however, and the extent of such differential familiarity remains to be ascertained more fully. In any case, even if it were found that the degree of familiarity between female kin was indeed greater than that between male and female kin, other evidence indicates that such differential familiarity is not sufficient to account for the present contrast between homosexual and heterosexual incest avoidance. First, one would expect a difference of degree between these two patterns, whereas it was found that homosexual females were never incestuous. Second, if differential familiarity was all that mattered, males and females would not necessarily differ in their attitude toward incest and we would expect them t o share the responsibility in initiating it. However, as mentioned previously, the available evidence suggests that males are more responsible than are females for the exceptions to incest avoidance. The second hypothesis is therefore unlikely. CONCLUSIONS 1. All sexually mature females in a captive group of Japanese macaques engaged in homosexual activity over four consecutive mating seasons. While all females had many non-kin partners, they systematically avoided sexual interactions with their mother, daughters, and sisters. 2. This pattern could not be fully explained in terms of kin not being simultaneously in estrus, kin avoiding affiliative interactions in general, or non-kin making use of the potential tension-reducing effect of estrus to affiliate exclusively with each other. 3. In marked contrast with such total absence of incest among homosexual females, heterosexual pairs were sometimes incestuous. 4. Assuming that female homosexuality reflects the reproductive strategy of females unconstrained by male influence, the present data point to the strong tendency of females to avoid incest and suggest that males are primarily responsible for the observed exceptions to that pattern. ACKNOWLEDGMENTS We thank Carole Gauthier, Michelle Girard, Alain Houle, FranCois Larose, and Ginette Primi for technical assistance and Daniel Perusse for critical reading of the manuscript and for suggesting the third alternative explanation to incest avoidance. We are grateful to Sarah Hrdy, Meredith Small, and Linda Wolfe for their constructive criticism of the manuscript. The University of Washington Primate Information Center is gratefully acknowledged for a bibliography on incest avoidance and homosexual behavior in nonhuman primates. This research was funded by the Natural Sciences and Engineering Research Council of Canada, the Fonds FCAR of the province of Quebec, and the Universite de Montreal. REFERENCES Abbott, D.H. Behavioral and physiological suppression of fertility in subordinate marmoset monkeys. AMERICAN JOURNAL OF PRIMATOLOGY 6:169-186,1984. Akers, J.S.; Conaway, C.H. Female homosexual behavior in Macaca mulatta. ARCHIVES OF SEXUAL BEHAVIOR 8:6380, 1979. Baxter, M.J.; Fedigan, L.M. Grooming and consort partner selection in a troop of Japanese monkeys (Macaca fuscata). ARCHIVES OF SEXUAL BEHAVIOR 8:445458,1979. Chapais, B. Matriline membership and male rhesus reaching high ranks in their natal troops. Pp. 171-175 in PRIMATE SOCIAL 182 1 Chapais and Mignault RELATIONSHIPS. R.A. Hinde, ed. Sunderland, Sinauer, 1983a. Chapais, B. Reproductive activity in relation to male dominance and the likelihood of ovulation in rhesus monkeys. BEHAVIORAL ECOLOGY AND SOCIOBIOLOGY 12:215-228, 198313. Chapais, B. Rank maintenance in female Japanese macaques: Experimental evidence for social dependency. BEHAVIOUR 104:41-59, 1988. Chapais, B.; Girard, M.; Primi, G. Non-kin alliances and the stability of matrilineal dominance relations in Japanese macaques. ANIMAL BEHAVIOUR, in press. Clark, T.; Mano, T. Transplantation and adaptation of a troop of Japanese monkeys to a Texas brushland habitat. Pp. 358-361 in CONTEMPORARY PRIMATOLOGY. S. Kondo, M. Kawai, A. Ehara, eds. Basel, Karger, 1975. Clutton-Brock, T.H.; Harvey, P.H. Evolutionary rules and primate societies. Pp. 195-237 in GROWING POINTS IN ETHOLOGY. P.P.G. Bateson, R.A. Hinde, eds. Cambridge, Cambridge University Press, 1976. Datta, S.B. The acquisition of dominance among free-ranging rhesus monkey siblings. ANIMAL BEHAVIOUR 36:754775, 1988. Dunbar, R.I.M.; Dunbar, P. SOCIAL DYNAMICS OF GELADA BABOONS. Basel. Karger, 1975. Eaton, G.G. Longitudinal studies of sexual behavior in the Oregon troop of Japanese macaques. Pp. 35-59 in SEX AND BEHAVIOR. T.E. Gill, D.A. Dewsbury, B.D. Sachs, eds. New York, Plenum, 1978. Ehardt, C.; Bernstein, I.S. Patterns of affiliation among immature rhesus monkeys (Macaca mulatta). AMERICAN JOURNAL OF PRIMATOLOGY 13:255-269, 1987. Enomoto, T. On social preference in sexual behavior of Japanese monkeys (Macaca fuscata). JOURNAL OF HUMAN EVOLUTION 7:283-293, 1978. Goodall, J. THE CHIMPANZEES OF GOMBE. PATTERNS OF BEHAVIOR. Cambridge, Belknap Press of Harvard University Press, 1986. Gouzoules, H.; Goy, R.W. Physiological and social influences on mounting behavior of troop-living female monkeys (Macaca fuscata). AMERICAN JOURNAL OF PRIMATOLOGY 5:39-49, 1983. Itani, J . A preliminary essay on the relationship between social organization and incest avoidance in non-human primates. Pp. 165-171 in PRIMATE SOCIALIZATION. F.E. Poirier, ed. New York, Random House, 1972. Koford, C.B. Rank of mothers and sons in bands of rhesus monkeys. SCIENCE 141: 356-357, 1963. Kummer, H. SOCIAL ORGANIZATION OF HAMADRYAS BABOONS. Chicago, University of Chicago Press, 1968. Lee, P.C. Sibship cooperation and competition among immature vervet monkeys. PRIMATES 28%-70, 1987. Loy, J. Estrous behavior of free-ranging rhesus monkeys (Macaca mulatta). PRIMATES 12:l-31, 1971. Missakian, E.A. Genealogical mating activity in free-ranging groups of rhesus monkeys (Macaca mulatta) on Cay0 Santiago. BEHAVIOUR 45:225-241, 1973. Murray, R.D.; Smith, E.O. The role of dominance and intrafamilial bonding in the avoidance of close inbreeding. JOURNAL OF HUMAN EVOLUTION 12:481-486, 1983. Packer, C. Inter-troop transfer and inbreeding avoidance in Papio anubis. ANIMAL BEHAVIOUR 27:l-36, 1979. Paul, A.; Kuester, J. Intergroup transfer and incest avoidance in semifree-ranging Barbary macaques (Macaca sylvanus) a t Salem FRG. AMERICAN JOURNAL OF PRIMATOLOGY 8:317-322,1985. Pereira, M.E. Effects of age and sex on intra-group spacing behaviour in juvenile savannah baboons, Papio cynocephalus cynocephalus. ANIMAL BEHAVIOUR 36: 184-204,1988. Pusey, A. Inbreeding avoidance in chimpanzees. ANIMAL BEHAVIOUR 28:543-552, 1980. Pusey, A,; Packer, C. Dispersal and philopatry. Pp. 250-256 in PRIMATE SOCIETIES. B.B. Smuts et al., eds. Chicago, University of Chicago Press, 1987. Sade, D.S. Inhibition of son-mother mating among free-ranging rhesus monkeys. SCIENCE AND PSYCHOANALYSIS 12:1838, 1968. Sade, D.S. A longitudinal study of social behavior of rhesus monkeys. Pp. 378-98 in THE FUNCTIONAL AND EVOLUTIONARY BIOLOGY OF PRIMATES. R. Tuttle, ed. Chicago, Aldine-Atherton, 1972. Savage, E.S.; Malick, C. Play and sociosexual behaviour in a captive chimpanzee (Pan troglodytes) group. BEHAVIOUR 60: 179-194, 1977. Scott, L.M. Reproductive behavior of adolescent female baboons (Papio anubis) in Kenya. Pp. 77-100 in FEMALE PRIMATES: STUDIES BY WOMEN PRIMATOLOGISTS. M.F. Small, ed. New York, Alan R. Liss, Inc., 1984. Smuts, B.B. SEX AND FRIENDSHIP IN BABOONS. New York, Aldine, 1985. Smuts, B.B. Sexual competition and mate Homosexual Incest Avoidance in Macaques I 183 choice. Pp. 385-399 in PRIMATE SOCIETIES. B.B. Smuts et al., eds. Chicago: University of Chicago Press, 1987. Symons, D. THE EVOLUTION OF HUMAN SEXUALITY. New York, Oxford University Press, 1979. Takahata, Y. The socio-sexual behavior of Japanese monkeys. ZEITSCHRIFT FUR TIERPSYCHOLOGIE 59:89-108,1982, Taylor, L.; Sussman, R.W. A preliminary study of kinship and social organization in a semi-free-ranging group of Lemur cattu. INTERNATIONAL JOURNAL OF PRIMATOLOGY 6:601-614,1985. Tilford, B. Seasonal rank changes for adolescent and subadult natal males in a free-ranging group of rhesus monkeys. INTERNATIONAL JOURNAL OF PRIMATOLOGY 3:483-490, 1982. Tokuda, K. A study on the sexual behavior in the Japanese monkey troop. PRIMATES 311-40, 1961-62. Trivers, R.L. Parental investment and sexual selection. Pp. 136-179 in SEXUAL SELECTION AND THE DESCENT OF MAN, 1871-1971. B. Campbell, ed. Chicago, Aldine, 1972. Westermarck, E. THE HISTORY OF HUMAN MARRIAGE. London, Macmillan, 1891. Wolfe, L. Behavioral patterns of estrous females of the Arashiyama West troop of Japanese macaques (‘Macaca fuscatu). PRIMATES 20525-534, 1979. Wolfe, L.D. Japanese macaque female sexual behavior: A comparison of Arashiyama East and West. Pp. 141-157 in FEMALE PRIMATES: STUDIES BY WOMEN PRIMATOLOGISTS. M.F. Small, ed. New York, Alan R. Liss, Inc., 1984. Wolfe, L.D. Sexual strategies of female Japanese macaques (Mucuca fuscuta). HUMAN EVOLUTION 1:267-275,1986.