最近の更新履歴 nkutsukake

全文

(1)

DOI: 10.1007/s10764-005-5308-4

Dominance Turnover Between an Alpha

and a Beta Male and Dynamics of Social

Relationships in Japanese Macaques

Nobuyuki Kutsukake1,2,3and Toshikazu Hasegawa1

Received March 15, 2004; revision June 15, 2004; 2nd revision July 30, 2004; accepted August 4, 2004

We report a case of turnover between an alpha (GN) and a beta male (R7) and its effects in a troop of provisioned Japanese macaques(Macaca fuscata fuscata) in Shiga-Heights, Nagano Prefecture, Japan. The aggression be- tween the 2 males was caused by the intrusion of GN towards the consort of R7. R7 received support from his brother and mother, and consequently de- feated GN. After the turnover,R7 attacked GN frequently, which may have functioned to stabilize the dominance relationship between them. Also, R7 selectively attacked females friendly to GN soon after the turnover. Although we never observed polyadic aggression among males during the stable domi- nance period, 20 cases of polyadic aggression occurred among the 6 highest- ranked males in the 2 days following the turnover, and one case occurred on the fourth day. R7 and GN formed stable conservative alliances for at- tacking subordinate males. Males that did not participate in the turnover began to form revolutionary coalitions to attack higher-ranking males, but they were thwarted by the conservative coalitions between the dominants. Mutualism was a plausible explanation for the patterns of coalition forma- tion because most of them were conservative with little associated cost. Seven females had a high proximity index (C-score) to GN before the turnover, but a significantly lower proximity index after the turnover. On the day of the

1Department of Cognitive and Behavioral Science, Graduate School of Arts and Sciences, The University of Tokyo.

2Department of Biological Science, Graduate School of Science, The University of Tokyo.

3JSPS Research Fellow.

4To whom correspondence should be addressed at 7-3-1 Hongo, Bunkyo-ku, Tokyo, Japan 113–0033; e-mail: kutsu@darwin.c.u-tokyo.ac.jp.

775

0164-0291/05/0800-0775/0C2005 Springer Science+Business Media, Inc.

(2)

turnover, 6 non-lactating females suddenly became receptive, suggesting that the turnover induced immediate receptivity in the females.

KEY WORDS: Japanese macaques; turnover between the alpha and beta male; coalitions and alliances; dominance struggle; friendships; receptivity.

INTRODUCTION

Many primate species living in a stable social group have a dominance relationship among group members. Occasionally, the dominance relation- ship changes as a result of severe aggression within the group (Ehardt and Bernstein, 1986; Gygax et al., 1997). In such cases, the aggressive turnover of a top-ranked individual, usually an alpha male, by members of the same group may be the most drastic social event in the society, because the top- ranked individual has a strong influence restricting subordinate behavior, such as access to resources (Keddy-Hector and Raleigh, 1992; Pusey & Packer, 1997; Silk, 1987; Walter and Seyfarth, 1987). However, only a few records detail agonistic turnovers between a top-ranked male and group member(s) in a stable group wherein the membership was not artificially disturbed, either in the wild or in captivity (Itoigawa, 1993; Mizuhara, 1964; Nakamichi et al., 1995; Nishida, 1983; Perry, 1998a,b; Sprague et al., 1996; de Waal, 1982, 1986; Witt et al., 1981) as a result of their unpredictability and the difficulty in observation. In addition, it seems likely that the reports on dominance turnover have included a substantial bias in the data because reporters in anecdotal studies did not use systematic observation methods.

During a research project on the social behavior of Japanese macaques (Macaca fuscata fuscata; Kutsukake, 2000; Kutsukake and Castles, 2001), we observed the whole process of agonistic turnover between the alpha male (GN) and beta male (R7). We describe how the turnover occurred and via systematic observations and data collected for other studies, we report on 3 aspects of the social dynamics after change of the alpha male. First, we report the strategy used by a male to maintain his dominance or challenge higher-ranking males. We not only focused on the interaction between the alpha and beta males whose dominance position changed but also exam- ined the dominance struggles among other males resulting from coalition formation, (i.e., coordinated aggression towards conspecifics). Coalitions represent cooperative behavior, and their occurrence has been confirmed in various primates (Harcourt and de Waal, 1992; Chapais, 1995; Cords, 1997). Coalitions may incur a substantial cost to the participants (Dunbar, 1988), and are termed alliances when they last for a long time and remain stable (Harcourt and de Waal, 1992). Males sometimes struggle for domi- nance position via coalitions during or following a dominance rank change

(3)

(Itoigawa, 1993; Mizuhara, 1964; Nakamichi et al., 1995; Perry, 1998a,b; Witt et al., 1981; de Waal, 1982; 1986; Nishida, 1983; Nishida and Hosaka, 1996). No previous researcher has systematically analyzed male-male coali- tions in Japanese macaques because, in general, fewer cases of coopera- tive bonds have been confirmed among male versus female primates (van Hooff and van Schaik, 1992, 1994), and because male Japanese macaques rarely form coalitions in the context of intra-group actions (Koyama, 1967; Watanabe, 1979; Grewal, 1980; Takahata, 1982). In our study group all 6 high-ranking males were from the lineage of the alpha female, which may be uncommon in male-dispersing species like macaques. Natal males from the alpha female lineage can postpone emigration to gain alpha-status in provi- sioned groups (Chapais, 1983), which provides an opportunity to examine the strategies of coalition formation, partner choice, and the influences of kinship in coalition formation among male Japanese macaques. Also, previ- ous studies showed special role of the alpha female in macaque society; she often influences dominance among males (Gouzoules, 1980; Chapais, 1983; Chapais and Lecomte, 1995; Nakamichi et al., 1995; Raleigh and McGuire, 1989). For example, Chapais (1983) observed a provisioned free-ranging group of rhesus macaques and reported that sons of an alpha female oc- cupied high-ranking positions with her support, a very similar process to matrilineal rank inheritance among females. As the situation in our study group is quite similar to that of Chapais (1983), we paid particular attention to the role of the alpha female during the dominance instability period.

Second, we investigated changes in heterosexual friendly relation- ships on turnover of the alpha male. Various studies of primates living in multimale-multifemale societies have confirmed special, affiliative rela- tionships between adult males and unrelated, anestrous females for non- sexual purposes (Smuts, 1999). The male-female relationships—friendships (Smuts, 1999)—are typically characterized by a high frequency of proxim- ity, frequent grooming, and agonistic support.

Given that a relationship can be described as investments that benefit the participants (Kummer, 1978), it is predictable that relationship quality changes when the dominance rank of the friendship partner changes. How- ever, only a few researchers have investigated the changes in male-female friendship with a turnover. Nakamichi et al. (1995) observed turnover be- tween an old alpha-male and a beta-male in Japanese macaques and found that the number of adult females in proximity to the new alpha male in- creased after the male gained the alpha position (Raleigh and McGuire, 1989: vervet monkeys Cercopithecus aethiops). Perry (1997, 1998b) ob- served replacement of the alpha male in wild white-faced capuchin monkeys Cebus capucinus and reported that the proximity scores of adult females and the previous alpha male decreased after his rank declined.

(4)

Third, we report a sudden change in receptivity of nonlactating fe- males soon after the turnover. Previous studies showed that female pri- mates do not always cycle regularly, and various social factors influence facultative female receptivity. For example, exposure to a novel male facili- tates the start of female receptivity in macaques (Wilson and Gordon, 1979; Okayasu et al., 1992). Female receptivity also can change after takeover of the group in hamadryas baboons (Colmenares and Gomendio, 1988; Swedell, 2000; Zinner and Deschner, 2000). Our case is unique, because the intra group change in male-dominance rank induced changes in female receptivity.

METHODS

Study Troop, Troop Composition, and Background

We examined Shiga A-1, a free-ranging troop of Japanese macaques in Shiga Heights, Nagano Prefecture, Japan. Demographic records—births, deaths, immigrations, and emigrations—dating to 1962 are available and all individuals are identifiable. The staff of the Jigokudani Monkey Park provi- sions the macaques 3–4 times daily and tourists are prohibited from feeding them. Japanese macaques have a clear mating and nonmating seasons. The mating season is, ca. from October to February at Shiga Heights, but most cases of mating occurred from October through December. The turnover occurred on 9 November 1998, during the mating season. The first author observed turnover process from a distance of 10 m, from which we could record the interactions in detail.

Kutsukake has intermittently observed the troop since October 1997, and has studied dominance and aggression among group members (Kutsukake, 2000; Kutsukake and Castles, 2001). Basically, observations were from 08:00 to 17:00 h. The troop comprised ca. 230 individuals, in- cluding 22 adult (>6 yr) males and 77 adult (>5 yr) females. The group contained 5 kin groups (T, N, K, R and M); among which, all the kin groups except M stemmed from one matriline (Kutsukake, 2000). All of the fe- males are basically named according to their kin group, e.g., MB is from kin group M. Because kin group T was a large kin group containing >40 adult females, we also used family names to indicate the subunits within kin group T, e.g., the alpha female and her daughter are the T(g) group. The matrilineal rank inheritance rule (Kawamura, 1958; Chapais, 1992) predicts that kin group M ranked below the other 4 kin groups. Kutsukake (2000) investigated the dominance relationships among 69 adult females >6 yr old in 1998. Based on the average dominance ranks of all females in each kin group, the dominance relationship among kin groups was T > M > N > R

(5)

>K. In particular, 2 M females (MB and MR) were dominant over nearly half of the females in kin group T (Figure 2 in Kutsukake, 2000).

Ten of the 22 adult males usually stayed in the center of the troop, and we determined the dominance order among them by observing retreat- supplant behaviors. GN, a healthy 22-yr-old, acquired the status of alpha male in December 1997, after the previous alpha male emigrated to a neigh- boring troop. He was the brother of the alpha female (TG: 24 yr). A beta male R7 (11 yr) was a son of TG and thus a nephew of GN. Seven highest- ranking males belonged to the matriline of TG. (Fig. 1). The stable domi- nance order among 7 high-ranking males before the turnover was GN > R7

>R4 (14 yr) > ZK (6 yr) > N (9 yr) > M (10 yr) > IN (7 yr). For simplicity we used male names that do not begin with T, though they were from kin group T(g). Males ranking below the seventh rank were from 2 kin groups: T [though not from the T(g) family] and middle-ranking N.

Several females unrelated to GN were frequently in proximity to him in the period since we started observations in October 1997. M kin-group members, in particular were most often in proximity to GN. The fact that kin group M remained near GN, had frequent grooming interactions with him, and received his support during aggressive interactions, probably explains why 2 females in M dominated the females expected to become

Fig. 1. Genealogy among the 6 highest-ranked males and the alpha female (TG). The numbers written below correspond to the year of birth, and the numbers above each male indicate his dominance rank before and after the turnover.

(6)

dominant based on matrilineal rank inheritance rules (Kawamura, 1958; Chapais, 1992; Chapais and Lecomte, 1995). According to Eishi Tokida (personal comm.) director of Jigokudani Monkey Park, the relationships were established before the start of our observations.

GN emigrated from the troop in September 1999, 10 mo after the turnover. However, it is unknown whether his emigration was related to the turnover Observations in May 2000 confirmed that GN was solitary.

Observation Methods

Kutsukake and Castles (2001) observed most of the group members daily and opportunistically recorded all aggressive interactions, which re- vealed the general characteristics of intra-group aggression.

Following the turnover, we used focal animal sampling (Altmann, 1974) to observe the previous alpha male (GN) for 10 days (total obser- vation: 72 h). Simultaneously, we conducted all occurrence sampling for male-male interactions among the 6 highest-ranked males (Altmann, 1974). Kutsukake stood where he could look around the whole area of the feed- ing site and observe their positions and behaviors. Whenever the distance among >2 of the 6 highest-ranked males became <10 m, we paid special attention to them. In general, it is impossible to conduct focal observations and all-occurrence sampling observations simultaneously. However, males usually stayed in an open area, where visibility was quite good, were inac- tive and did not change their position frequently, and were immobile during the mating season. Male dyads, except R7 and GN, kept their distance from each other and rarely associated. As a result of the favorable observation conditions, Kutsukake was able to conduct both focal and all occurrence sampling simultaneously and to observe ≥4 males for > 98% of the obser- vation time. Therefore, it is highly likely that we observed all polyadic inter- actions among them while they were at the feeding site. However, we were unable to observe the macaques on the morning of the third day after the turnover because they remained in the mountains and did not come to the feeding site. Only one polyadic interaction involved IN, the seventh-ranked male (fig. 1), which received redirected aggression from the third-ranked male, R4. Other males did not participate in the polyadic interactions among the 6 highest-ranked males. This indicates that not all high-ranking adult males from T(g) participated in the dominance competition.

We checked dominance relationships among males via the sampling ad libitum (Altmann, 1974) by daily observations of supplant-retreat be- haviors. The dominance relationship among males did not change, except the turnover between GN and R7.

(7)

To investigate the friendly females for GN and R7, we used proximity as an index of friendship (Kurland, 1977; Cords, 1997). Kutsukake recorded adult females proximate (0–0.5 m, 0.5–1 m, and 1 − <5 m) to GN or R7 every hour for 3 observation periods:

Preturnover (mating season): 19–25 October and 7–8 November 1998, resulting in 17 scans for GN and 15 scans for R7.

Posturnover (mating season): 9–18 November 1998, resulting in 39 scans for GN and 40 scans for R7.

Stable (non-mating season): 38 days from March–September 1999, when GN emigrated from the troop, resulting in 76 and 77 scans, respec- tively.

For all 3 periods, we collected data only during periods of no artificial feeding. We include no datum from December to February in the analysis because heavy snow caused the macaques to concentrate by a hot spring to thermoregulate. As an index of affiliative relationships between females and males, we used a composite proximity score (C-score), which consid- ered the proximity level of a female to a male (Smuts, 1999; Perry 1997, 1998b). The following equation derived the scores:

C-score = (P0−0.5 m)/1 + (P0.5−1 m)/3 + (P1−5 m)/12

wherein PA−Bm is the proportion of proximity from A meter to B meter. Each weight figure corresponds to the proportion of the middle point of each range (0.25 for 0∼0.5 m, 0.75 for 0.5 ∼ 1 m, and 3 for 1 ∼ 5 m). To allow the scores to range between 0 (never ≤ 5 m) and 1 (always ≤ 0.5 m), we multiplied the proportions by the arbitrary value 4, and used 1, 3, and 12 as weighted values. For preturnover period data, we calculated an aver- age C-score for each female in proximity to GN and R7 at least once. The number of scans is small because of the sudden, unpredictable occurrence of the turnover and the restricted data collection due to the heavy snow. However, we believe that the estimate of the proximate relationship is re- liable because the distribution of the C score is not continuous and most of the females had extremely low C scores (fig. 2c, f), while the C scores of several unrelated females are higher than those of other females (fig. 2a, b, d). We defined those females as affiliative females of GN or R7.

Jigokudani Monkey Park staff and Kutsukake recorded reproductive states of adult females during the mating season. We determined receptivity by the criteria of copulation and the presence of a seminal plug.

Definitions

Aggression is indicated by facial threat, threat vocalization, confronta- tion while staring at each other, chasing, lunging, grasping, and biting. If

(8)

Fig. 2. Changes in the C-score of each female before the turnover, after the turnover, and during the dominance-stable period. (a) Females from kin group M that were friendly with GN; (b) 3 females from other kin groups; (c) 3 females from the T(g) family group and other females (N = 21) that were in proximity to GN at least once during the pre-turnover, post- turnover or stable period; (d) females from the T(e) family group that were friendly with R7; (e) 3 females that increased their proximity to R7 after the turnover; (f) females (N = 17) that were in proximity to R7 at least once during the pre-turnover, post-turnover or stable period.

two aggressive interactions were separated by >1 min, they were counted as separate interactions. Support is the intervention of a male towards one of a pair of males in conflict. A coalition is any interaction in which ≥2 males jointly attack ≥1 male. Coalitions were classified as revolutionary, conservative, or bridging coalitions (Chapais, 1995). Revolutionary coali- tions form between subordinates against a dominant, (e.g., alpha vs. beta and gamma). Conservative coalitions are between dominants against a sub- ordinate, e.g., alpha and beta vs. gamma. Bridging coalitions are between

(9)

2 individuals against a middle-ranked individual, e.g., alpha and gamma vs. beta). Japanese macaques sometimes solicit aid from other males by repeat- edly turning the head quickly from an opponent to the coalition partner. Therefore, we also recorded the occurrence of the head flag display and its target.

Data Analysis

First, we present the characteristics of aggression between adults dur- ing the period of stable dominance. The data are from Kutsukake and Castles (2001), who observed 543 cases of aggression during ca. 630 h of observation between June 1998 and March 2000. (aggressive interactions observed during the period of the present study were not included in the dataset). We also deleted aggression data that involved juveniles in order to present a general picture of aggression between adults in the troop. We paid particular attention to the frequency of polyadic aggression.

With regard to the behavior of males following the turnover, we in- vestigated the number of polyadic interactions and the number of supports, coalitions formed, and solicits-for-aid in dyad levels as follows. If A sup- ported B, which had confronted C in one polyadic interaction. We recorded that A had supported B once. When the direction and the initiation of the support was unclear, we excluded the data from the analysis. We counted the number of coalitions formed in each dyad by ignoring the direction and initiation of agonistic support. For instance, we counted one coalition in an A-B dyad. When 3 males (A, B, C) simultaneously attacked male (D), we recorded that a coalition was formed in 3 dyads: A-B, B-C, and A-C. Coali- tion partners often changed in one polyadic interaction, in which case, we counted one coalition when it was formed once in a dyad. For example, if A and B cooperatively attacked C, but then A and C formed a coalition against D in the middle of the battle, we recorded one coalition in the A-B and A-C dyads, and support from A to C.

To investigate the effect of kinship, rank differences, and age differ- ences on the frequency of coalition formation, we conducted a matrix cor- relation test (Hemelrijk 1990), which is a revised version of the Mantel test (Schnell et al., 1985) that considers interindividual variation in behavior frequency. We conducted 5000 permutations. Since the paternity of the 6 highest-ranked adult males was unknown, we considered only maternal kin- ship in the analysis, i.e., r = 0.25 for brothers, r = 0.125 for uncle-nephew pairs, and r = 0.0625 for distant kin such as granduncle-grandnephew pairs (Chapais et al., 1997). One could argue that it is meaningless to test the rela- tionship between maternal kinship and the frequency of coalition formation because all of the males were from T(g) and related. Chapais et al. (1997)

(10)

investigated the relatedness threshold for nepotism, e.g., agonistic support, for female Japanese macaques and found that nepotistic behavior drops be- yond grandmother-granddaughter and sibling dyads, i.e., r = 0.25 (Kurland, 1977; Belisle and Chapais, 2001). Although the relatedness threshold for male Japanese macaques has not been investigated, it is reasonable to as- sume that males have a similar relatedness threshold of nepotism or a higher relatedness threshold, given that males interact less with closely-related in- dividuals than females do, which validates the analysis of the relationship between maternal kinship and coalition-formation frequency because the maternal relatedness among the 6 highest-ranked males varied widely.

Via the Wilcoxon signed-rank test, we compared the mean C-score of GN- and R7-affiliative females before and after the turnover. Because the baseline for male-female proximity differs between the mating and non- mating seasons, we did not compare pre-or post-turnover conditions with stable-period conditions. The significance level is p < 0.05.

RESULTS Background

We observed 407 dyadic and polyadic aggressive interactions and clas- sified them by the combination of sexes between the opponents Table I. During the stable dominance period, dyadic aggression between adult males (n = 7) is far less than between adult male and adult female (n = 61) or between adult females (n = 302). After considering that Shiga A- 1 contained fewer adult males (n = 22) than adult females (n = 77), we found that the number of aggressive interactions in the 3 categories of

Table I. Dyadic and polyadic aggression during the dominance stable period

Sex combination # cases

Dyadic aggression M vs. M 7

M vs. F 61

F vs. F 302

Polyadic aggression M & M vs. M 0 M & M vs. F 0 M & F vs. M 12 M & F vs. F 9 F & F vs. M 5 F & F vs. F 8

3F vs. F 2

M & 2F vs. M 1 Note.M: male; F: female.

(11)

sex combination differed from the expected numbers (Chi-squared test; χ2(2)=66.2, P < 0.0001), i.e., the number of female-female aggressions is significantly higher than the number of male-female and male-male aggres- sions (Fisher’s exact probability test, Ps < 0.001). Of 37 cases of polyadic ag- gressive interactions, none involved 3 males or male-male coalitions attack- ing adult females; the absolute number of dyadic and polyadic (male-male coalition vs. single male) aggressive interactions among males or male-male coalitions was quite low during the period of stable dominance.

Before the turnover, 27 females were in proximity to GN at least once. Visual inspection of the C-score distribution suggests that 7 females that were not from T(g) clearly had higher C-scores than those of other females (Figure 2a, b, c), and they were the only females whose C-scores exceeded the mean (0.12) of the C-score distribution. Four (MB, MG, MR, and MK) were from M, while the other 3 females (TW, TM, and NM) were from other kin groups (Table II). This result coincides with and reinforces our previous observation that M females formed special affiliative relationships with GN (Kutsukake, 2000). NM was in receptivity during these periods and was guarded by GN. On the other hand, the 4 M females and TW and TM were not receptive.

Before the turnover, 17 females were in proximity to R7 at least once. Visual inspection of the C-score distribution suggests that among them, 3 clearly had higher scores than the others, and they were the only fe- males whose C-scores exceeded the mean (0.03) of the C-score distribution (Figure 2d,f). All were from T(e), which is a branch of T, and they were not receptive before the turnover.

Table II. Identity and reproductive condition of the females

Affiliative Kin Reproductive Bore in Latest

male Name Age group condition 1999 infant

GN MB 24 M Menopausal? No 1993

MR 19 M Active No 1994

MG 5 M Active Yes Nulliparous

MK 6 M Active No Nulliparous

TW 17 T Lactating No 1998

TM 17 T Lactating No 1998

NM 8 N Active Yes 1997

R7 Te5 13 T(e) Active Yes 1997

Te1 7 T(e) Lactating No 1998

Te3 15 T(e) Lactating No 1998

TW (see above)

KR 15 K Active Yes 1998 (dead)

TK 19 T Active No 1994

Related females TG 24 T(g) Active Yes 1997

Tg1 7 T(g) Active No 1998

Tg3 5 T(g) Lactating No 1998

(12)

Turnover Between the Alpha and Beta Males

At 1000 h on 9 November, GN interfered between R7 and his con- sorting female (Ngn: 24 yr) and attacked her. R7 rushed GN, and they began to tug and bite each other, which lasted for ca. 3 min. This ag- gressive interaction was unusually loud. Most members of the troop ob- served the interaction, and nearly 30 individuals surrounded the conflict within a distance of 20 m. During the aggression, all 7 GN-affiliative fe- males were present but did not participant in the aggression or support GN. R4 (the third-ranked male and R7’s brother) and TG (the alpha fe- male and R7’s mother) joined the aggression and supported R7 by biting GN. In the course of the conflict, TG threatened and lunged at R4, even though both had jointly supported R7 against GN. R4 stopped attacking GN. R7 then bit GN for about 1 min. When injured GN escaped from R7, he first showed the bared-teeth expression, which signals subordina- tion, to R7, indicating that the dominance relationship between them had reversed. R7 and most of the individuals that had surrounded the conflict, including R4 and TG, followed GN. TG then moved in front of R7 and performed a facial threat, as though she were intervening in the aggres- sion, but R7 ignored her. GN lay down and became immobile at 10:04h, and R7 sat ca. 5 m from him. After the aggression concluded, R7 was in proximity to GN, and attacked him 15 times in 2 h. The alpha female, TG, and her daughter (Tg3) stayed with GN and R7, and groomed GN. GN’s friendly females, except for TW, were found near GN and R7. How- ever, TW stayed with GN and R7. TW presented and copulated with R7 in the presence of GN 11 times within 1 h of the turnover. Because, we concentrated on observing male-male interactions, we can not specify the location of the other GN-affiliative females except for MR and TM, which sat ca. 30 m from GN. R7 approached and attacked them successively 9 times within 34 min after the turnover. There were several adult females other than MR, TM, and TG near GN, but R7 did not attack them. In ad- dition, aggression by R7 directed against MR or TM is nil in the aggres- sion data during the stable dominance condition (Kutsukake and Castles, 2001), rejecting the possibility that R7 attacked them frequently irrespec- tive of the turnover. Furthermore, no persistent attack towards females other than GN-affiliative females occurred after the turnover, which facts suggest that R7’s attacks were not random but highly selective against MR and TM.

GN received at least 8 wounds that bled, possibly from R7’s canines. We assume that the injures did not cause severe injury to GN because he was able to walk normally the following day. However, he remained rel- atively immobile, did not feed for the remainder of the day, and did not

(13)

engage in sexual activity for 3 days. R7 received 5 wounds that bled, but he seemed to have no difficulty performing his daily activities. R4 and TG were not wounded.

Dyadic Aggressive Interactions Among Adult Males

R7 attacked GN 35 times in the 9 days following the dominance turnover (fig. 3a). Twenty-six aggressive interactions (74%) occurred on the first day. Aggression was always unidirectional and included physical con- tact and in most cases (33/35 cases) GN showed the bared-teeth facial ex- pression. We observed no reconciliation—a postconflict affiliation between the opponents (Kutsukake and Castles, 2001; Aureli et al., 2002)—between them. The first explicit affiliation, grooming, between them occurred 6 days after the turnover (15 Nov).

Except for the GN-R7 dyad, the 6 highest-ranked adult males were seldom near each other, and we observed neither dyadic aggression nor grooming interactions among them.

Polyadic Aggression Among the Six Highest-Ranked Males After the turnover, we observed 21 polyadic interactions among the 6 highest-ranked males. Twenty of 21 polyadic interactions occurred within 2 days after the turnover (Figure 3b). In polyadic aggressive interactions, the mean number of adult males that were involved is 3.38 (min: 3; max: 5; median: 3). No polyadic interaction among males involved physical contact. The most frequent pattern (in 15 polyadic interactions) was GN at- tacking a lower-ranked male, which occurred mainly within 1.5 days of the turnover (Figure 3b). Aggression was directed toward R4 9 times; GN so- licited R7 all 9 times and R7 supported GN 5 times. GN attacked ZK 4 times, for which GN solicited R7 3 times and R7 supported GN twice. GN attacked N 3 times, for which GN solicited R7 all 3 times, but R7 never sup- ported GN. Accordingly, GN became quite aggressive toward the lower- ranking males and enlisted support from R7.

Seven of 21 polyadic interactions included aggression directed towards a dominant male. GN was the target of aggression in 4 polyadic interactions (by R4 once, ZK once, N-M coalition once, and M once), R7 in one polyadic interaction (by N-M coalition), R7-GN coalition in one polyadic interaction (R4-ZK coalition once), and R4-ZK coalition in 1 polyadic interaction (by N-M coalition once).

In 3 cases of relatively simple polyadic interactions in which GN was attacked by a lower-ranking male, R7 supported him in reaction to his

(14)

Fig. 3. (a) Frequency of aggression of R7 towards GN following the turnover per h. G indi- cates the first observed grooming interaction between R7 and GN. (b) Number of polyadic aggression involving the 6 highest-ranking males. The dashed line indicates the number of cases of aggression directed from GN to lower-ranking male(s). In both (a) and (b), observa- tions could not be conducted in the morning of November 11, 1998.

solicitation of support, and they won the encounter (against R4, ZK, and M once each). Also, GN solicited R7 against R4 once, but R7 did not sup- port him. R4 retreated.

In rather complex polyadic interactions involving 4 males, lower- ranking males formed revolutionary coalitions:

(15)

Example 1 (10:56 h, 9 Nov): R4 threatened GN. ZK then sided with R4 and threatened GN. R7 approached and sided with GN, (which did not solicit support from R7), and confronted the R4-ZK coalition. GN chased R4, and R4 chased (redirected towards) ZK.

Example 2 (13:23 h, 9 Nov): N and M approached and sided with each other, and moved towards R7, which was threatened by the N-M coalition, but ignored them. After that, the N-M coalition moved to GN and threat- ened GN. GN ignored them. M dissolved the coalition and moved away from the confrontation. N continued to threaten GN alone, but eventually moved away.

Example 3 (15:25 h, 12 Nov): N and M approached and sided with each other and moved towards R4. R4 was threatened by the N-M coalition. ZK and TG approached the confrontation and sided with R4. M dissolved the coalition and fled from the R4-ZK-TG coalition. N bit TG but was chased away by R4 and ZK.

The examples indicate that the males that did not participate in the turnover (ZK, N, and M) also began a dominance challenge and that both the reduced-in-rank GN and other males (R7 and R4) became targets of the dominance challenges by lower-ranking males.

Coalition Formation and Soliciting for Aid Among Males

Regarding coalition formation, we were unable to determine the di- rection in all polyadic interactions since individuals who formed a coalition approached simultaneously and began the aggression, e.g., N-M coalition in examples 2 and 3. Table III is a support-receiver matrix of cases in which the identity of the supporter was apparent. In 11 of 12 cases, the dominant was supported against a subordinate. Table IV shows the number of coalitions formed, not considering the direction of the agonistic support. We observed 16 conservative coalitions and 4 revolutionary coalitions, while, there was

Table III. Agonistic support among 4 high-ranking males

Receiver R7 GN R4 ZK

R7 0 6D 1D 0

Supporter GN 1D 0 1D 0

R4 0 1D 0 0

ZK 0 0 1D + 1S 0

S : the support to a subordinate male against a dominant male. D: the support to a dominant male against a subordinate male.

(16)

Table IV. Coalition formation among the 6 highest-ranking males

R7 GN R4 ZK N M

R7 9C 2C 1C 0 0

GN 3C 0 0 0

R4 1C+1R 0 0

ZK 0 0

N 3R

M

C: the conservative coalition. R: the revolutionary coalition.

no bridging coalition, which suggests that revolutionary coalition is rare. Of the 16 conservative coalitions, 9 were between GN and R7, indicating that they formed a stable alliance against lower-ranking males.

Via a matrix correlation test, coalition formation is not related to ma- ternal relatedness (Figure 4: TauKr = 0.773, P = 0.06), age differences (TauKr = 0.202, P = 0.35) or rank differences (TauKr = −0.337, P = 0.82).

The head-flagging display occurred in 17 polyadic interactions, 21 times among 5 dyads (Table V). All times, except for one from R7 to GN, were directed from a subordinate to a dominant male. GN displayed most fre- quently to R7 (16 times), and succeeded in obtaining his support 5 times (31%). GN never solicited support from another male. R7 displayed to GN once against R4 and GN supported him. R7 did not respond to solicitations from any male other than GN, though R4 and ZK each displayed to R7 once. GN supported both of R4’s solicitations, even though GN frequently attacked him with support from R7.

Fig. 4. Number of coalitions formed among the 6 highest-ranking males. Relationships with (a) maternal kinship, (b) rank difference, (c) age difference.

(17)

Table V. Solicit for aid among the 4 highest-ranking males Direction # head

(hierarchy flagged

Initiator Target up/down) display # success

GN R7 up 16 5

R7 GN down 1 1

R4 R7 up 1 0

R4 GN up 2 2

ZK R7 up 1 0

Changes in the C-Score for Each Female Before and After Turnover

We compared the C-score for each affiliative female to GN before and after the turnover, and found a significant C-score decrease after the turnover (Figure 2a, b; Wilcoxon signed-rank test: N = 7, T = 0, p = 0.016). After the turnover, only the alpha-female and her daughters had high C- scores (Figure 2c) and frequently groomed GN. Other females were rarely in proximity to GN. However, in the next non-mating season (1999), 4 M females also showed the highest C-scores for GN. This suggests that sepa- ration by M females from GN after the turnover was temporal and did not last until the nonmating season. Other females were not proximate to GN in the next nonmating season (Figure 2b, c). Regarding the females affiliated with R7, we could not compare the C scores before and after the turnover because of the small sample size (N = 3; Figure 2d).

For each scan during the pre- and post-turnover periods and the sta- ble period, we calculated the total C-score of all females for GN and R7 (Figure 5), and compared the total C-score before and after the turnover via the Mann-Whitney U test. For GN, the total C-score decreased signif- icantly (Npre=17, Npost=39, U = 129.0, P < 0.0001), indicating that the many females that had surrounded GN before the turnover separated after the turnover. For R7, no significant difference showed before or after the turnover (Npre=15, Nafter =40, U = 262.0, P = 0.55).

Change of Receptivity in Females

On the turnover day, at least 6 of 52 nonlactating adult females with no previous sign of sexual receptivity began to show signs of receptivity. Via simulation, we calculated the likelihood of beginning of sexual recep- tivity in them on the same day. The null hypothesis is that the simultaneous start of receptivity could occur in 6 females in their menstrual cycles. The length of the cycle in nonpregnant females (31.8 ± 11.8 days) is based on

(18)

Fig. 5. Mean (+SE) of the total C-scores for GN and R7 before the turnover, after the turnover, and during the dominance-stable period.

Takahata’s (1980) research on another provisioned troop in Arashiyama, Japan. We created an artificial data set of 52 female cycles. The data set has a normal distribution and the mean and SD matches Takahata’s (1980) data. We refer to the length of each cycle randomly generated as Mi. The possibility that X of Y females began receptivity on a certain day can be calculated by the following function (Dunbar, 1988, 1999; Nunn, 1999):

Y!

X! × (Y − X)! ×

X i=1

(1/Mi) ×

Y−X j=1

(1 − 1/Mj)

The simulation revealed that the probability of 6 females becoming recep- tive on the same day (0.004%) was less than the significant level. R7 cop- ulated with 2 of the females. In the following year (1999), 14 females gave birth in the troop; however, 2 females did not.

TW also became receptive on the turnover day, but she is not included in the analysis because she was lactating then. TW showed a sudden C-score increase for R7 (Figure 2e) and copulated with him. TK and KR also had an increased C-score for R7 (Figure 2e) after the turnover and copulated with him. Because TK and KR had been receptive when the turnover occurred, their increased C-scores for R7 were not caused by the start of receptivity. However, the 3 females did not birth the next year.

(19)

DISCUSSION

Turnover and Coalitionary Aggression Among Males

In this study, mating conflict resulted in the turnover between the al- pha and beta males, and the turnover occurred with severe aggression in- volving a high cost of physical injury. Although the absolute number of dyadic and polyadic aggressive interactions was low among high-ranking males during periods of dominance stability, because of the demography of the group (fewer adult males than adult females), there was a sudden in- crease in aggression among males, both dyadic, i.e., aggression from R7 to GN, and polyadic. The increased frequency of aggression during the domi- nance instability period concurs with previous observations on chimpanzees and white-faced capuchin monkeys (de Waal, 1982; Perry 1998ab). In this study, the plausible function of dyadic aggression was to stabilize the new dominance hierarchy, i.e., R7 attempted to stabilize his dominance over GN by frequently attacking him soon after the turnover (Figure 3a) even though GN had already shown the formal signal of submission soon af- ter the turnover. Conversely, GN might have attempted to maintain his beta position with the help of R7 in polyadic interactions, despite fre- quent aggression from R7. Also, GN was the only male from which R7 solicited aid against R4. Furthermore, R7 used GN as a conservative al- liance partner to defend his alpha position from challenges by subordi- nates, even against his own brother R4. Indeed, R7 reacted to 31% of GN’s solicitations for aid and supported him. The data suggest that R7 and GN formed a conservative alliance, despite the fact that R7 harassed GN within the dyadic dominance relationship. A conservative alliance between alpha and beta males also occurred in chimpanzees (de Waal, 1982).

Males also used aggression in order to destabilize the dominance rank after the turnover. Not only the males involved in the turnover (R7, GN, R4) but also the subordinates (ZK, N, M) contributed to the aggression. In addition, both the reduced-in-rank GN and other males, such as R7 and R4, became targets of dominance challenges. This suggests that the turnover and temporal dominance instability stimulated the males to try to increase their ranks. Coalitionary dominance struggles among males have been re- ported only anecdotally or in nonsystematic observations in macaques by Itoigawa (1993), Mizuhara (1964), Nakamichi et al. (1995), and Witt et al., (1981), and a quantified analysis failed to show that male macaques use coalitions to enhance their social rank in male bonnet macaques Macaca ra- diata(Silk, 1993). Regular occurrence of male-male coalition formation to improve social rank—political coalitions—in dominance competitions had

(20)

been reported only in chimpanzees by de Waal, (1982), Nishida, (1983), and Nishida and Hosaka, (1996).

One distinct characteristic of the polyadic dominance competitions was that males avoided the cost of engaging in severe aggression: polyadic ag- gression never developed into physical aggression, even when the lower- ranking males had numerical superiority when attacking the dominant male (N and M attacking GN or R7: example 2), most coalitions were conserva- tive and incurred little cost for the coalition partners, and lower-ranking males stopped forming revolutionary coalitions once the higher-ranking males had formed conservative coalitions (examples 1 and 3). Via series of experimental studies, Chapais and colleagues who created subgroups of female Japanese macaques to induce rank reversals, noted the importance of alliances with kin (Chapais, 1988) and non-kin (Chapais et el., 1991, 1994; Chapais & St-Pierre, 1997) in the process of rank inheritance. Subordinate female Japanese macaques challenged the dominant individual only when the power asymmetry was obviously in favor of the subordinates (Chapais, 1992). Thus, it appears that female Japanese macaques follow a minimum- risk strategy, i.e., they avoid and reduce the risk of aggression when the expected it is high. The female behaviors seem similar to the low-cost strat- egy by males in this study group, which indicates that males also followed a minimum-risk strategy during dominance instability.

Theoretically, cooperation can evolve through several mechanisms, such as mutualism, reciprocal altruism, and kin selection (Dugatkin, 1997). Although we had few cases of coalition formation and agonistic support to analyze the ultimate factor quantitatively, it is likely that the behav- iors were based on mutualism because most of the coalitions were con- servative with little associated cost. Moreover, stabilizing the dominance relationship offers immediate benefits to the participants of the coalition. Several researchers have suggested that male coalitions may be mutualistic (Bercovitch, 1988; Noe, 1992; Watts, 1998; Widdig et al., 2000).

Behavior of the Alpha Female, TG

TG was the only female that participated to the dominance turnover (supporting R7 against GN). Supporting a more closely related individ- ual (son) than a more distant one (brother) may be reasonable for TG. However, TG’s aggression towards R4 during the cooperative aggression towards GN is difficult to understand because TG did not need to attack R4 to defeat GN. TG also threatened R7 after the submission signal by GN to R7, though TG had supported R7 in the defeat of GN. Further, TG was the only female that participated in the polyadic interactions among

(21)

the males (example 3), which suggests that the alpha female held a spe- cial position in the group, and possibly influenced male dominance ranks to some extent (Gouzoules, 1980; Chapais, 1983; Chapais and Lecomte, 1995; Nakamichi et al., 1995; see also Raleigh & McGuire, 1989). Given the influ- ence of the alpha female, it may be beneficial for natal males that are born to a high-ranking matriline to postpone emigration and to attain the dom- inant position in their natal group with the help of the females (Chapais, 1983).

In this study, TG increased proximity to deposed alpha male GN rather than proximity to R7. Nakamichi et al. (1995) observed a similar case of turnover in another troop of Japanese macaques and found that females, e.g., the alpha female, increased their proximity to the new alpha male after turnover. In Shiga A-1, the alpha female was a sister of GN, which have reduced the total C-score for R7, as compared to the situation ob- served by Nakamichi et al. (1995). The relatedness between TG and all of the highest-ranking males may have made the influence of the alpha female much stronger in this study group.

Changes in Females: Fragile Friendship and Flexible Receptivity Our study showed that proximity between friendly adult females and the deposed alpha male temporarily disappeared following turnover. More- over, the total C score to GN per scan decreased after the turnover. The re- sults agree with Perry’s (1997, 1998b) observations that proximity scores of adult females and defeated previous alpha males decreased after the male’s defeat. In contrast, the proximate relationship of R7 and adult females did not appear to be affected by the turnover, except for the 3 receptive females that increased proximity to R7, which disagrees with previous studies that showed increases in female proximity to the new alpha male (Nakamichi et al., 1995; Raleigh and McGuire, 1989).

We supposed that avoidance of aggression from R7 was one of the fac- tors explaining the short-term change in relations between GN and friendly females for the following 3 reasons. First, frequent aggression from R7 to GN occurred soon after the turnover. The aggression may have pre- vented friendly females from accessing GN, because interacting with the target of aggression incurs the cost of becoming involved in the aggres- sion. In Japanese macaques, previously uninvolved third parties affiliate significantly less with victims of aggression as compared to affiliation un- der normal conditions (Aureli et al., 1993). Second, females with a C- score increase after the turnover were related to R7. In general, related females showed tolerance, irrespective of dominance position. Finally, R7

(22)

attacked GN-affiliative females soon after the turnover. Because the fe- males were not spatially close to GN at the time of attack, it appears that R7 targeted the females selectively. The attacks also suggest that R7 recognized the affiliative relationships between GN and his friendly females.

Not all females friendly to GN, behaved similarly. Notably, TW ac- tively copulated with and maintained proximity to R7 after the turnover, and she was not attacked by R7. TW became receptive on the day of the turnover. TW’s behavior may have functioned actively to cultivate interac- tions with R7 and thus avoid R7’s attacks.

In addition, the long-term effects of the dominance turnover on male- female friendship varied; M females resumed their proximate relationship to GN, but TM and TW did not. For M females, it may have been ben- eficial to remain friendly with the beta male GN because the M was his- torically subordinate to the other kin groups and the strong bond with GN might have improved their dominance rank over other matrilineal groups (Kutsukake, 2000). Conversely, TM and TW were from the high-ranking T kin group and a loss of benefits resulting from ending friendship with GN may have been less important for them.

We found that 6 adult females became receptive on the day of the turnover. Simulation showed a significance level indicating that this would not occur coincidentally. Although several studies have shown that the fe- male receptive condition changes in reaction to exposure to a new male (Wilson and Gordon, 1979; Okayasu et al., 1992), no study has shown that changes in intra-group social situations induced immediate receptivity in female primates. The benefits of the onset of receptivity by social stimula- tion within a group are unclear. It is possible that the physiological mech- anism that led to the onset of receptivity when exposed to a new male was a response to the unusual intra-group dynamic event, i.e., the dominance reversal between the alpha and beta-male. However, we cannot draw con- clusions about the adaptive significance of the phenomenon from our data, and we need more studies to reveal the adaptive significance of immediate receptivity.

Generality of the Study and Importance of N == 1 Observation Because our results are based on a single case we hesitate to conclude that the behavior we observed represents or is typical of the species. For ex- ample, regarding male-male interaction, there is little evidence that male- male coalitions play an important role in dominance relationships among males in unprovisioned wild Japanese macaques (Furuichi, 1983, 1985).

(23)

Male-male coalitions may reflect a high frequency of aggression and an- tagonistic relationships among males under provisioned conditions (Hill, 1999). Moreover, there is much dissimilarity with the study by Nakamichi et al.(1995), in which the males did not form a coalition and struggle for rank. The high relatedness among the highest-ranking males, which is un- usual in a society with female philopatry and male dispersal, also might have facilitated the formation of coalitions among males. Therefore, dominance turnover and its influences vary with the case.

What kind of general biological conclusions can we draw from case studies of a rare event? As Chapais (1985) and de Waal (1991) noted, they are valuable to provide new insight into behavioral flexibility and social strategy. As long as researchers use systematic observation methods when observing rare cases, the results can lead to new hypotheses, which can be tested in controlled conditions. For example, experimental manipula- tion of group composition can address questions on the power dynamics of the dominance relationship among females (Chapais, 1992) and male- female friendship in Japanese macaques (Chapais and Lecomte, 1995). Such paradigms are very useful to test the phenomena that we have re- ported. In addition, captive studies can reveal hormonal changes under con- ditions of varying access to males (Mitsunaga et al., 1994), which is useful to reveal the physiological aspects of flexible female receptivity. Therefore, we emphasize the usefulness of single-case studies based on systematic ob- servations, like this one, to provide new testable questions.

ACKNOWLEDGMENTS

We thank Duncan Castles and Ichiro Tanaka for support over the course of this study. We especially thank Eishi Tokida, Sogo Hara, Haruo Takefushi, Toshio Hagiwara and other staffs of Jigokudani Monkey Park for permitting the study and supporting every stage of my fieldwork. C. Hemelrijk kindly provided the program of Matrixtester, for which we are grateful. Reviewers provided critical comments on the manuscript, for which we are grateful. The study was supported by JSPS Research Fellow- ships and The 21st Century COE Program (The Center for Evolutionary Cognitive Sciences at the University of Tokyo).

REFERENCES

Altmann, J. (1974). Observational study of behavior: sampling methods. Behaviour 49: 227– 265

Aureli, F., Cords, M., and van Schaik, C. P. (2002). Conflict resolution following aggression in gregarious animals: a predictive framework. Anim. Behav. 64: 325–343

(24)

Aureli, F., Veenema, H. C., van Panthaleon van Eck, C. J., and van Hooff, J. A. R. A. M. (1993). Reconciliation, consolation, and redirection in Japanese macaques (Macaca fus- cata). Behaviour 124: 1–21.

Belisle, P., and Chapais, B. (2001). Tolerated co-feeding in relation to degree of kinship in Japanese macaques. Behaviour 138: 487–509.

Bercovitch, F. B. (1988). Coalitions, cooperation and reproductive tactics among adult male baboons. Anim. Behav. 36: 1198–1209

Chapais, B. (1983). Matriline membership and male rhesus reaching high ranks in the natal troops. In: Hinde, R. A. (ed). Primate social relationships: an integrated approach. Oxford, Blackwell Scientific Publications. pp. 171–175.

Chapais, B. (1985). An experimental analysis of a mother-daughter rank reversal in Japanese macaques (Macaca fuscata). Priamtes 26: 407–423.

Chapais, B. (1988). Experimental matrilineal inheritance of rank in female Japanese macaques. Anim. Behav. 36: 1025–1037.

Chapais, B. (1992). The role of alliances in social inheritance of rank among female primates. In: Harcourt, A. H., and de Waal, F. B. M. (eds). Coalitions and alliances in humans and other animals, New York, Oxford University Press, pp. 29–59.

Chapais, B. (1995). Alliances as a means of competition in primates: evolutionary, develop- mental, and cognitive aspects. Yearbook of Physical Anthropology 38: 115–136. Chapais, B., Girard, M., and Primi, G. (1991). Non-kin alliances, and the stability of matrilineal

dominance relations in Japanese macaques. Anim. Behav. 41: 481–491.

Chapais, B., and Lecomte, M. (1995). Induction of matrilineal rank instability by the alpha male in a group of Japanese macaques. Amer. J. Primatol. 36: 299–312

Chapais, B., Prud’Homme, J., and Teijeiro, S. (1994). Dominance competition among siblings in Japanese macaques: constraints on nepotism. Anim. Behav. 48: 1335–1347.

Chapais, B., and St-Pierre, C. G. (1997). Kinship bonds are not necessary for maintaining ma- trilineal rank in captive Japanese macaques. Int. J. Primatol. 18: 375–385.

Chapais, B., Gauthier, C., Prud’Homme, J., and Vasey, P. (1997). Relatedness threshold for nepotism in Japanese macaques. Anim. Behav. 53: 1089–1101.

Colmenares, F., and Gomendio, M. (1988). Changes in female reproductive condition follow- ing male take-overs in a colony of hamadryas and hybrid baboons. Folia Primatologica 50: 157–174.

Cords, M. (1997). Friendships, alliances, reciprocity and repair. In: Whiten, A., and Byrne, R. W. (eds). Machiavellian intelligence II –extentions and evaluations-. Cambridge, Cam- bridge University Press, pp. 24–49.

Dugatkin, L. E. (1997). Cooperation among animals. Oxford, Oxford University Press. Dunbar, R. I. M. (1988). Primate social systems. London & Sydney, Croom Helm.

Dunbar, R. I. M. (1999). Male mating strategies: an optimal foraging problem. In: Kappler, P. (ed) Primate males–causes and consequences of variation in group composition. Cam- bridge, Cambridge University Press.

Ehardt, C. L., and Bernstein, I. S. (1986). Matrilineal overthrows in rhesus monkey groups. Int. J. Primatol.7: 157–181.

Furuichi, T. (1983). Interivdividual distance and influence of dominance on feeding in a natural Japanese macaque troop. Primates 24: 445–455.

Furuichi, T. (1985). Inter-male associations in a wild Japanese macaque troop on Yakushima Island, Japan. Primates 26: 219–237.

Gouzoules, H. (1980). The alpha female: Observations on captive pigtail monkeys. Folia Pri- matologica. 33: 46–56.

Grewal, B. S. (1980). Social relationships between adult central males and kinship groups of Japanese monkeys at Arashiyama with some aspects of troop organization. Primates 21: 161–180.

Gygax, L., Harley, N., and Kummer, H. (1997). A matrilineal overthrow with destructive ag- gression in Macaca fascucularis. Primates 38 149–158.

Haucourt, A. H, and de Waal, F. B. M. (1992). Coalitions and Alliances in Humans and Other Animals. Oxford, Oxford University Press.

(25)

Hemelrijk, C. K. (1990). Models of, and tests for, reciprocity, unidirectionality and other social interaction patterns at a group level. Anim. Behav. 39: 1013–1029.

Hill, D. A. (1999). Effects of provisioning on the social behaviour of Japanese and rhesus macaques: implications for socioecology. Primates 40: 187–198.

van Hooff, J. A. R. A. M., and van Schaik, C. P. (1992). Cooperation in competition: the ecology of primates bonds. In: Harcourt, A. H., and de Waal, F. B. M. (eds). Coalitions and alliances in humans and other animals. New York, Oxford University Press. pp. 357– 389.

van Hooff, J. A. R. A. M., and van Schaik, C. P. (1994). Male bonds: Affiliative relationships among nonhuman primate males. Behaviour 130: 309–337.

Itoigawa, N. (1993). Social conflict in adult male relationships in a free-ranging group of Japanese monkeys. In Mason, W. A. and Mendoza, S. P. (eds). Primate social conflict, Albany, State University of New York Press, pp. 145–169.

Kawamura, S. (1958). The matriarchal social order in the Minoo-B group -A study on the rank system of Japanese macaque-. Primates 1: 149–156. (in Japanese)

Keddy Hector, A., and Raleigh, M. J. (1992). The effects of temporary removal of the alpha male on the behavior of subordinate male vervet monkeys. Amer. J. Primatol 26: 77– 87.

Koyama, N. (1967). On dominance rank and kinship of a wild Japanese monkey troop in Arashiyama. Primates 8: 189–216.

Kummer, H. (1978). On the value of social relationships to nonhuman primates: A heuristic scheme. Soc. Scie. Info. 17: 687–705.

Kurland, J. A. (1977). Kin selection in the Japanese monkey. Contributions to Primatology 12. Kutsukake, N. (2000). Matrilineal rank inheritance varies with absolute rank in Japanese

macaques. Primates. 41: 321–335.

Kutsukake, N., and Castles, L. D. (2001). Reconciliation and variation in postconflict stress in Japanese macaques (Macaca fuscata fuscata): testing the integrated hypothesis. Animal Cognition. 4: 259–268.

Mitsunaga, F., Nozaki, M., and Shimizu, K. (1994). Suppressed copulatory behavior and ovar- ian function in lactating Japanese monkeys (Macaca fuscata fuscata) during the mating season. Primates 35: 79–88.

Mizuhara, H. (1964). Social changes of Japanese monkey troops in the Takasakiyama. Pri- mates5: 27–52.

Nakamichi, M., Kojima, Y., Itoigawa, S., Imakawa, S., and Machida, S. (1995). Interactions among adult males and females before and after the death of the alpha male in a free- ranging troop of Japanese macaques. Primates 36: 385–396.

Nishida, T. (1983). Alpha status and agonistic alliance in wild chimpanzees (Pan troglodytes schweinfurthii). Primates 24: 318–336.

Nishida, T., and Hosaka, K. (1996). Coalition strategies among adult male chimpanzees of the Mahale Mountains, Tanzania. In McGrew, W. C., Marchant, L. F., and Nishida, T., (eds). Great Ape Societies. Cambridge, Cambridge University Press. pp. 114–134.

Noe, R. (1992). Alliance formation among male baboons: Shopping for profitable partners. In: Harcourt, A. H., and de Waal, F. B. M. (eds). Coalitions and alliances in humans and other animals. New York, Oxford University Press. pp. 285–321.

Nunn, C. L. (1999). The number of males in primate social groups: a comparative test of the socioecological model. Behav. Ecol. Sociobiol. 46: 1–13.

Okayasu, N. (1992). Prolonged estrus in female Japanese macaques (Macaca fuscata yakui) and the social influence on estrus: with special reference to male intertroop movement. In: Itoigawa, N., Sugiyama, Y., Sackett, G. P., and Thompson, R. K. R. (eds). Topics in Primatology, Vol 2 Behavior, Ecology, and Conservation, Tokyo, University of Tokyo Press. pp. 163–178.

Pusey, A. E., and Packer, C. 1997. The ecology of relationships. In Krebs, J. & Davies, N.B. (eds). Behavioural Ecology, Fourth edition, Blackwell, Oxford. pp. 254–283.

Perry, S. (1997). Male-female social relationships in wild white-faced capuchins (Cebus capuci- nus). Behaviour. 134: 477–510.

(26)

Perry, S. (1998a). Male-male relationships in wild white-faced capuchins, Cebus capucinus. Behaviour135: 139–172.

Perry, S. (1998b). A case report of a male rank reversal in a group of wild white-faced ca- puchins (Cebus capucinus). Primates 39: 51–70.

Raleigh, M. J., and McGuire, M. T. (1989). Female influences on male dominance acquisi- tion in captive vervet monkeys, Cercopithecus aethiops sabaeus. Anim. Behav. 38: 59– 67.

Schnell, G. D., Watt, D. J., and Douglas, M. E. (1985). Statistical comparison of proximity matrices: Applications in animal behaviour. Anim. Behav. 33: 239–253.

Silk, J. B. (1987). Social behavior in evolutionary perspective. In: Smuts, B. B., Cheney, D. L., Seyfarth, R. M., Wrangham, R. W., Struhsaker, T. T. (eds). Primate Societies. Chicago, University of Chicago Press. pp. 318–329.

Silk, J. B. (1993). Does participation in coalitions influence dominance relationships among male bonnet macaques? Behaviour 126: 171–189.

Smuts, B. B. (1999). Sex and friendship in baboons - a version with a new preface-. Cambridge, Harvard University Press.

Sprague, D. S., Suzuki, S., and Tsukahara, T. (1996). Variation in social mechanisms by which males attained the alpha rank among Japanese macaques. In: Fa, J. E., and Lindburg, D. G., (eds). Evolution and ecology of macaque societies, Cambridge, Cambridge University Press, pp. 444–457.

Swedell, L. (2000). Two takeovers in wild hamadryas baboons. Folia Primatologica 71: 169– 172.

Takahata, Y. (1980). The reproductive biology of a free-ranging troop of Japanese monkeys. Primates21: 303–329.

Takahata, Y. (1982). Social relations between adult males and females of Japanese monkeys in the Arashiyama B troop. Primates 23: 1–23.

de Waal, F. B. M. (1982). Chimpanzee Politics -Power and Sex among Apes. London, The Johns Hopkins University Press.

de Waal, F. B. M. (1986). The brutal elimination of a rival among captive male chimpanzees. Ethology and Sociobiology7: 237–251.

de Waal, F. B. M. (1991). Complementary methods and convergent evidence in the study of primate social cognition. Behaviour. 118: 297–320.

Walters, J. R., Seyfarth, R. M. (1987). Conflict and cooperation. In: Smuts, B. B., Cheney, D. L., Seyfarth, R. M., Wrangham, R. W., Struhsaker, T. T. (eds). Primate Societies. Chicago, University of Chicago Press. pp. 306–317.

Watanabe, K. (1979). Alliance formation in a free-ranging troop of Japanese macaques. Pri- mates20: 459–474.

Watts, D. P. (1998). Coalitionary mate guarding by male chimpanzees at Ngogo, Kibale Na- tional Park, Uganda. Behav. Ecol. Sociobiol. 44: 43–55.

Widdig, A., Streich, W. J., and Tembrock, G. (2000). Coalition formation among male Barbary macaques (Macaca sylvanus). Amer. J. Primatol. 50: 37–51.

Wilson, M. E., and Gordon, T. P. (1979). Sexual activity of male rhesus monkeys introduced into a heterosexual group. Amer. J. Phys. Anthrop. 50: 515–524.

Witt, R., Schmidt, C., and Schmidt, J. (1981). Social rank and Darwinian fitness in a multimale group of Barbary macaques (Macaca sylvana Linnaeus, 1758) -Dominance reversals and male reproductive sucess-. Folia Primatol 36: 201–211.

Zinner, D., and Deschner, T. (2000). Sexual swelling in female hamadryas baboons after male take-overs: “deceptive” swellings as a possible female counter-strategy against infanticide. Amer. J. Primatol. 52: 157–168.

Fig. 1. Genealogy among the 6 highest-ranked males and the alpha female (TG). The numbers
Fig. 1. Genealogy among the 6 highest-ranked males and the alpha female (TG). The numbers p.5
Fig. 2. Changes in the C-score of each female before the turnover, after the turnover, and
Fig. 2. Changes in the C-score of each female before the turnover, after the turnover, and p.8
Table I. Dyadic and polyadic aggression during the

Table I.

Dyadic and polyadic aggression during the p.10
Table II. Identity and reproductive condition of the females

Table II.

Identity and reproductive condition of the females p.11
Fig. 3. (a) Frequency of aggression of R7 towards GN following the turnover per h. G indi-
Fig. 3. (a) Frequency of aggression of R7 towards GN following the turnover per h. G indi- p.14
Table III. Agonistic support among 4 high-ranking males

Table III.

Agonistic support among 4 high-ranking males p.15
Fig. 4. Number of coalitions formed among the 6 highest-ranking males. Relationships with
Fig. 4. Number of coalitions formed among the 6 highest-ranking males. Relationships with p.16
Table IV. Coalition formation among the 6 highest-ranking males R7 GN R4 ZK N M R7 — 9C 2C 1C 0 0 GN — 3C 0 0 0 R4 — 1C+1R 0 0 ZK — 0 0 N — 3R M —

Table IV.

Coalition formation among the 6 highest-ranking males R7 GN R4 ZK N M R7 — 9C 2C 1C 0 0 GN — 3C 0 0 0 R4 — 1C+1R 0 0 ZK — 0 0 N — 3R M — p.16
Table V. Solicit for aid among the 4 highest-ranking males

Table V.

Solicit for aid among the 4 highest-ranking males p.17
Fig. 5. Mean (+SE) of the total C-scores for GN and R7 before the turnover,
Fig. 5. Mean (+SE) of the total C-scores for GN and R7 before the turnover, p.18

参照