International Journal of Primatology

, Volume 29, Issue 1, pp 225–243

Evaluating Dominance Styles in Assamese and Rhesus Macaques

Authors

    • Department of PsychologyUniversity of Tennessee
  • Irwin S. Bernstein
    • Department of PsychologyUniversity of Georgia
Article

DOI: 10.1007/s10764-008-9236-y

Cite this article as:
Cooper, M.A. & Bernstein, I.S. Int J Primatol (2008) 29: 225. doi:10.1007/s10764-008-9236-y

Abstract

Researchers have suggested that several types of agonistic and affiliative behavior covary as a set of species-specific traits, and have used the term dominance style to describe the covariation. We compared measures of dominance style between a group of Assamese macaques (Macaca assamensis) and a group of rhesus macaques (M. mulatta), though kinship information was unknown. Assamese and rhesus female-female dyads each showed a low proportion of counter aggression and a low conciliatory tendency, suggesting that they have despotic social relationships. They also showed a despotic pattern on several other types of agonistic and affiliative behavior, such as approach outcomes and grooming distributions, which is consistent with the covariation of dominance style traits. Assamese male-male dyads showed relatively high levels of reconciliation and counter aggression versus other macaque males portrayed in the literature, suggesting that Assamese males have a tolerant dominance style. Insofar as macaque dominance style depends on the behavior of females, we suggest that Assamese macaques, like rhesus macaques, have despotic social relationships, which contrasts with evidence of a strong correlation between phylogeny and dominance style in macaques. Further, our results indicate that strong male bonding and tolerant dominance relationships among males are independent of female dominance style. Lastly, some measures of agonistic behavior, such as rate of aggression or proportion of bites, are likely altered in competitive environments and thus are not useful indicators of dominance style.

Keywords

counter aggressiondespoticdominance styleMacaca assamensisMacaca mulattareconciliation

Introduction

Researchers have long considered macaques aggressive primates, and have recognized species-specific differences in aggression directed toward humans. More than 50 yr ago, when Bernstein first began to work with them, rhesus macaques (Macaca mulatta) already had a reputation for being nearly intractable in the laboratory. Though some would cower and grimace on first approach by humans, others would vocally threaten, grab, and bite handlers immediately. Even rhesus monkeys hand-reared from birth would attack all but a few close associates, and bite if provoked. Other nonhuman primates were often far more manageable in the laboratory, and some individuals would tolerate familiar human caregivers and occasionally show positive social interest in people. For example, bonnet macaques (Macaca radiata) are much more social and tractable than pigtailed macaques (M. nemestrina) which, in turn, are far more easily handled than rhesus macaques (Kaufman and Rosenblum 1966). In fact, people in Southeast Asia have widely used pigtailed monkeys as coconut pickers and have often tied them up near their homes.

More recently, investigators have reported specific differences in intragroup social behavior and have suggested that some types of agonistic and affiliative behavior covary as a set of traits. Investigators originally coined the term dominance style to describe species-specific patterns of agonistic behavior (de Waal and Luttrell 1989), though today they use it more broadly to refer to the covariation of traits related to conflict and conflict management (Flack and de Waal 2004). Macaque species with a despotic dominance style have highly asymmetrical patterns of agonistic behavior, low rates of reconciliation, and a strong kin bias in affiliation (Aureli et al. 1997; Castles et al. 1996; de Waal and Luttrell 1989; Thierry 1985), whereas species with the opposite pattern of traits are considered tolerant or relaxed. Researchers have sometimes used egalitarian, tolerant, and relaxed inconsistently, but we use tolerant restrictively and reserve egalitarian for species that lack decided and stable dominance relationships. Dominance style in macaques is a graded series, whereby species exist along a continuum with despotic at one end and tolerant at the other (Thierry 2000). Other aspects of social organization may vary with the more traditional measures of dominance style. In general, macaques with a tolerant dominance style often show frequent exceptions to a strict inheritance of dominance rank, low rates of severe biting, tolerance around limited resources, little agonistic behavior in response to the approach of others, maternal tolerance for infant handling, and triadic male-infant interactions (Thierry 2000).

Researchers have proposed several models to explain consistent interspecific variation in social organization. Socioecological models suggest that the relative levels of intra- and intergroup competition influence female social relationships (Sterck et al. 1997). In this view, dominance styles represent adaptations by females and their kin that improve their access to resources and increase their reproductive success. Accordingly, a tolerant dominance style in macaques might be the product of selection pressures that promote collective action against predators, other groups of conspecifics, or potentially infanticidal males.

Researchers have proposed several models to explain why social organization may deviate from current selection pressures. Some have suggested that phylogenetic inertia precludes ready adaptation to local ecological pressures and that macaque dominance styles relate more to phylogeny than to local ecology (Matsumura 1999; Thierry et al. 2000). The interconnections between dominance style traits may act to constrain evolutionary change, and single agonistic traits may have self-organizing effects on others (Hemelrijk 1999; Thierry 1990). Further, sets of agonistic traits may represent evolutionary stable strategies that do not necessarily match ecological conditions (Matsumura and Kobayashi 1998). Conversely, some types of agonistic and affiliative behavior may reflect learned adaptations to a specific habitat, would not be constrained by phylogeny or past selection pressures, and would be poor indications of a species-specific dominance style. Thus, where Assamese macaques (Macaca assamensis) have begun to live at temples, along roadsides and in rural areas, they may converge on some of the behavioral adaptations of rhesus monkeys living under similar ecological conditions. Whereas Wilson (1975) invoked the multiplier effect to predict that genetic contributions would produce profound influences on social organization, Kummer (1971) suggested that because social organization was the aspect of behavior furthest removed from the genes, the relationship between phylogeny and social organization should be tenuous at best.

Assamese macaques, like most macaques, live in multimale-multifemale groups and have well established linear dominance hierarchies (Bernstein and Cooper 1999; Fooden 1982). Assamese macaques are members of the sinica lineage, and based on phylogenetic relatedness are expected to have a tolerant dominance style. Little information is available on the behavioral ecology of Assamese macaques and a dominance style prediction based on the socioecology model is not possible. In a previous study on Assamese macaques at a temple site, we found that female-female dyads had low levels of counter aggression and reconciliation, and suggested that females have a despotic dominance style (Cooper and Bernstein 2002). However, we have also reported that males frequently groom and mount one another, reconcile fights, and show a high proportion of counter aggression (Bernstein and Cooper 1999; Cooper and Bernstein 2000, 2002). Thus, the behavior of males appears to complicate the dominance style of Assamese macaques. In the present study, we compared dominance styles of a troop of Assamese macaques and a troop of rhesus macaques, each living at a temple site in Assam, India. If phylogeny is the primary influence on dominance style we would expect Assamese macaques to have a more tolerant dominance style than that of rhesus macaques. If some dominance style traits are, in fact, behavioral adaptations to local conditions we would expect that troops of rhesus macaques and Assamese macaques living under similar conditions might show similarities in certain dominance style traits.

Methods

Habitat and Subjects

Rhesus macaques are more widely distributed than Assamese macaques are, but they overlap with Assamese macaques from Nepal, through northeast India, and into western Thailand. Though rhesus and Assamese macaques are broadly sympatric in the region and live in similar habitats including moist deciduous forest, we did not observe the 2 species living side by side. We found Assamese macaques in forests, near farm fields, along roadsides, and at temple sites, but they were not as commensal with humans as rhesus macaques were.

Our group of Assamese macaques comprised 64 individuals including 11 adult males, 14 adult females, 3 subadult males, 29 juveniles, and 7 yearlings. We estimated age based on body size and the eruption of canines in males and did not have kinship information. The group lived on the grassy area around Tukeswari temple and on the steep boulder-strewn hill behind the temple, which was sparsely forested and surrounded by farmland. A second, smaller temple was at the peak of the 110-m hill. Priests fed the monkeys daily at both temples, and tourists often observed and fed the monkeys at the lower temple. In addition to offerings from priests and tourists, the monkeys foraged on the hill and raided adjacent rice fields. We occasionally saw another Assamese macaque troop in the rice fields and on the far side of the hill, but they did not overlap with the study group on the temple grounds.

The rhesus study group had 22 individuals including 2 adult males, 6 adult females, 1 subadult male, 9 juveniles, and 4 yearlings, and we did not have kinship information. The group lived at the Basistha temple compound, which included several buildings and a large stream, and in the reserve forest behind the temple which consisted primarily of a Sal (Shorea robusta) monoculture. Ritualized feedings by priests were not as regular at Basistha temple as at Tukeswari temple, but many more tourists visited Basistha. The monkeys received offerings from tourists, foraged in trash left behind by them and on undergrowth in the reserve forest, and occasionally raided shops near the temple. At least 2 other troops of rhesus macaques lived in the reserve forest behind the temple, with ranges that overlapped with the study group’s home range, but they seldom appeared at the temple complex. We also often saw 3 solitary males at the temple. Though the 3 males interacted among themselves, affiliative interactions with the study group were exceedingly rare and limited to the lone subadult male and an older juvenile male. In contrast, no Assamese male was peripheral, as each affiliated with ≥1 high-ranking male and several resident females.

The Assamese macaque group was 3 times larger than the rhesus group. The age distribution was similar for the 2 groups, but the sex ratio was not. In the Assamese macaque group 39% of individuals were adult and 45% were juvenile, versus the rhesus group, in which 36% of individuals were adult and 41% were juvenile. The adult sex ratio (males: females) was 1:1.3 in the Assamese macaque group and 1:3.0 in the rhesus group.

Observation Protocol

We collected observations on the Assamese macaques between 0700 and 1700 h from October to February, for a total of 1008 h. Our observations overlapped with the Assamese macaque’s mating season, which occurs from August to November. We began observations on the rhesus group immediately after the Assamese macaque study, and collected data between 0700 and 1700 h from March to May, for a total of 264 h. Our observations overlapped with the rhesus macaque birth season because theirs is similar to that of Assamese macaques. Cooper, Bernstein, and a local research assistant collected data on Assamese macaques and reached 90% reliability on focal samples. Cooper and the research assistant collected data on the rhesus group.

For each group, we recorded the occurrence and direction of agonistic behavior, grooming, and male-infant interactions ad libitum. We also conducted 4 h of focal sampling on each adult individual in each study group (24, 10-min samples), noting the date and time of each sample. During focal samples, we recorded all affiliative and agonistic interactions involving the subject, all approaches to ≤1 m, the identity of the initiator and recipient, and the time when the interaction occurred. Affiliative behavior included allogroom, embrace, contact (including manual touch, passive touch, and huddle), hold (for yearlings only), play, muzzle contact, lip-smack, genital touch, and mount (including hip touch). We recorded affiliative interactions with yearlings for adult males only. Agonistic behavior included silent bared-teeth display, lip-grin, supplant, avoid, flee, open-mouth threat, push, grab, lunge, charge, chase, manual contact aggression, bite, and severe bite. We recorded all agonistic responses individually for third-party interactions, indicating the direction of support. Though the silent-bared teeth display is affiliative in some macaque species (Petit and Thierry 1992), rhesus macaques use it as a submissive signal (Preuschoft 2004), and Assamese macaques appear to do so as well (Bernstein and Cooper 1999; personal observation). Finally, before the start of each focal sample, we recorded the identity of each adult and subadult individual in view, its distance from the focal subject, and the group activity. We defined distances as ≤5 m, 5–25 m, and >25 m and group activity as resting, feeding, or traveling. We used information on individual distances and group activity to select matched-control (MC) samples for reconciliation analysis.

We collected 10-min postconflict (PC) samples after all aggressive interactions that involved a charge or more intense forms of aggression. At the start of PC samples we recorded the distance between opponents, but otherwise the observational protocol for PC samples was the same as for focal samples. During PC samples, we were interested primarily in the first affiliative contact between former opponents. The PC sample began immediately after the last agonistic response in the conflict. If aggression restarted in ≤2 min, we extended the observation to have a 10-min PC sample. We conducted PC sampling on adult individuals and restricted them to fights between adults or between adults and subadults. We identified the 2 main opponents in polyadic fights, which involved aggression among ≥3 individuals and may have included juveniles. The main victim was the individual initially attacked, and the main aggressor was the one that attacked most intensively or for the greatest duration. We selected one of the main opponents as the focal subject. We gave priority as focal subject to the individual for which we had the least amount of data. We also attempted to sample each subject an equal number of times as the aggressor and victim but could not achieve complete counterbalancing.

In most studies of captive primates, researchers collect MC samples 24 h after the corresponding PC sample at a similar time of day, though they are usually delayed if the focal subject is involved in an aggressive interaction before the start of the MC. In field studies, however, it is often more efficient to select an appropriate MC sample from baseline data (Aureli 1992). We compared PC samples with MC samples chosen from a pool of focal samples. We selected an appropriate MC sample by matching group activity and the distance between subjects in PC and focal samples. When multiple matches were available, we selected the focal sample closest to the date of the PC sample.

Analysis

We constructed dominance hierarchies from the direction of all submissive interactions —silent bared-teeth display; supplant; outcome of aggression— for the Assamese and rhesus macaque study groups. We calculated the linearity index h′, which is based on Landau’s index but corrected for unknown relationships and the proportion of interactions in the less common direction (Singh et al. 2003). To indicate further the extent to which agonistic interactions were asymmetric within dyads, we calculated a directional inconsistency index for the proportion of all submissive interactions in the less frequent direction within dyads (de Waal 1977).

With focal sample data, we calculated hourly rates for aggression, approaches, grooming, and male-infant contact. Because the 2 groups were of different sizes, it is possible that hourly rates were influenced by the number of potential partners. The effect of group size, however, is not linear; therefore, it is not easily corrected for. Thus, we supplemented hourly rates of behavior with analyses on mean proportions for several types of behavior. Specifically, we calculated the mean proportion of bites, counter aggression, aiding, positive and negative approach outcomes, total grooming, grooming directed toward each sex class, grooming directed up the dominance hierarchy, and male-infant triadic interactions. We based behavioral measures on similar studies on dominance style (Castles et al. 1996; de Waal and Luttrell 1989). We defined counter aggression as instances of aggression of any kind to which the recipient responded with aggression of any kind. We defined negative approach outcomes as approaches resulting in silent bared-teeth display, pushing, grabbing, withdrawal by the approach recipient, or aggression from the approach recipient. We excluded approaches resulting in aggression by the approach initiator. We defined positive approach outcomes as approaches resulting in grooming, embracing, huddling, copulation, or other forms of affiliative contact. We defined triadic male-infant interactions as 2 adult males simultaneously contacting an infant, or each other, while one holds an infant. We defined severe biting as a potentially or actually damaging bite sustained for >5 s and accompanied by head-shaking, pinning the opponent to the ground, or injury. We also measured the grooming distribution by calculating a Shannon-Weaver heterogeneity index (H′) and corrected for group size with Buzas and Gibson’s (1969) evenness index. Using mean proportions of behavior also allowed us to combine focal sample data with data obtained ad libitum for each type of behavior except approaches, for which only focal data were available.

For analysis of reconciliation we used the PC-MC method, comparing the minute during which former opponents initiated affiliative contact in PC and MC samples (de Waal and Yoshihara 1983). We considered PC-MC pairs attracted, dispersed, or neutral depending on whether affiliative contact occurred sooner in the PC sample, sooner in the MC sample, or at the same time, respectively. If affiliative contact did not occur in either sample, we considered the PC-MC pair neutral. Then, to test for the occurrence of reconciliation, we compared the proportion of attracted PC-MC pairs to the proportion of dispersed PC-MC pairs. For comparisons between study groups, we report the corrected mean conciliatory tendency, which equals (attracted pairs – dispersed pairs)/(total number of PC-MC pairs) * 100 (Veenema et al. 1994).

Because the rhesus group had 2 males, and thus only 1 male-male dyad, statistical comparisons between Assamese and rhesus male-male dyads lacked sufficient power and we do not report them. However, we report statistical comparisons for interactions between males and other group members. Statistical tests used included Wilcoxon matched-pairs, Mann-Whitney U, and sign tests. For Wilcoxon tests we report the T value and exact probability when n ≤ 15. All statistical tests were 2-tailed, α level was 0.05, and we report means ± SE.

Results

Dominance

We recorded 2691 submissive interactions among adult Assamese macaques. Mixed-sex and single-sex dominance hierarchies show high indexes of linearity (Table I). Among adults, only 1.1% of interactions within dyads were directed in the less common direction. Submissive interactions in male-male, female-female, and male-female dyads are also highly asymmetric (Table I). We recorded 515 submissive interactions among rhesus macaques. Similar to Assamese macaques, mixed-sex and single-sex dominance hierarchies showed high linearity indexes (Table I). Submissive interactions also are highly asymmetric. In fact, male-male and female-female dyads had no interaction directed in the less common direction (Table I).
Table I

Dominance hierarchies

 

Macaca assamensis

Macaca mulatta

Corrected Landau’s index (h′)

Adults

0.98

1.0

Male-male

0.94

1.0

Female-female

0.89

1.0

Directional inconsistency index

Adults

30/2691 (1.1%)

12/515 (2.3%)

Male-male

5/694 (0.7%)

0/35 (0%)

Female-female

9/661 (1.4%)

0/266 (0%)

Male-female

16/1336 (1.2%)

12/214 (5.6%)

Dyads indicate actors followed by recipients.

Aggression

We calculated rates of aggression from focal sample data, and found no statistically significant difference between Assamese and rhesus macaques for each sex class (Table II). We combined data obtained ad libitum and focal sample data for our analysis of the mean proportion of aggressive interactions between adults with bites, aiding, and counter aggression. In the Assamese macaque group, we recorded 1001 aggressive interactions (mean = 40.0, SE = 12.0, range = 0–278), and in the rhesus group we recorded 262 aggressive interactions (mean = 32.7, SE = 8.4; range = 5–72). There is no significant difference in the mean proportion of bites between Assamese and rhesus macaques for any sex class (Table II). Also, for Assamese macaques we recorded 6 severe bites (0.60% of aggressive interactions), and for rhesus macaques we recorded 1 severe bite (0.38% of aggressive interactions). The mean proportion of aiding is not significantly different between species (Table II). Though Assamese females aided one another more often than rhesus females did, the difference is not statistically significant (U = 8.0, p = 0.064).
Table II

Agonistic behavior

 

Macaca assamensis

Macaca mulatta

U

p

N

Mean ± SE

N

Mean ± SE

Total aggression (hourly rate)

Adults

25

0.34 ± 0.10

8

0.27 ± 0.09

96.5

ns

Male-male

11

0.28 ± 0.11

2

0.0 ± 0.0

Female-female

14

0.11 ± 0.05

6

0.17 ± 0.09

36.5

ns

Male-female

11

0.30 ± 0.08

2

0.37 ± 0.13

7.5

ns

Female-male

14

0.04 ± 0.02

6

0.06 ± 0.03

34.5

ns

Bites (%)

Adults

23

7.3 ± 1.7

8

5.9 ± 2.8

83.5

ns

Male-male

10

2.5 ± 1.2

2

0.0 ± 0.0

Female-female

12

11.6 ± 3.6

6

12.5 ± 7.9

31.5

ns

Male-female

11

8.9 ± 2.8

2

3.0 ± 1.3

10.0

ns

Female-male

10

1.2 ± 0.9

6

1.8 ± 1.9

30.0

ns

Aiding (%)a

Adults

22

22.4 ± 4.9

8

15.1 ± 2.8

86.0

ns

Male-adult

10

28.4 ± 7.0

2

11.8 ± 5.3

6.0

ns

Female-adult

12

17.3 ± 6.8

6

16.2 ± 3.5

27.0

ns

Male-male

10

76.5 ± 6.6

2

4.2 ± 4.2

Female-female

8

88.9 ± 6.4

5

57.5 ± 16.6

8.0

.064

Counter aggression (%)b

Adults

25

16.8 ± 4.4

8

13.7 ± 3.9

90.0

ns

Male-male

11

25.4 ± 8.6

2

0.0 ± 0.0

Female-female

13

6.3 ± 2.6

6

2.5 ± 1.2

30.5

ns

Male-female

9

70.8 ± 11.8

2

26.6 ± 9.8

2.0

ns

Female-male

14

4.9 ± 1.4

6

17.9 ± 5.3

17.0

.038

Dyads indicate actors followed by recipients.

aWe calculated the mean proportion of aiding males initiated and females initiated as a proportion of the total amount of aggression they directed to adults. We calculated the mean proportion of aiding males directed to other males, and females directed to other females, as a proportion of the amount of aiding they initiated.

bCooper and Bernstein (2002) previously reported a portion of the data on counter aggression in Assamese macaques.

We previously reported proportions of counter aggression for the group of Assamese macaques (Cooper and Bernstein 2002); however, the values in this study differ slightly because we restricted our analysis to adults whereas we previously included subadults and older juveniles. The mean proportion of counter aggression is not significantly different between species (Table II); however, females show a greater proportion of counter aggression against males in rhesus macaques than in Assamese macaques (U = 17.0, p = 0.038).

Reconciliation

In the rhesus group, we collected 67 PC-MC pairs on 7 adults (mean = 9.6, SE = 1.0, range = 6–13). Sixty-one PC-MC pairs are from dyadic interactions, and the remaining 6 PC-MC pairs include the main opponents from polyadic interactions. Former opponents made affiliative contact in 11.2% (± 3.9) of PC samples and in 4.6% (± 2.3) of MC samples (Wilcoxon test: T = 1, n = 7, p = 0.035). Likewise, the proportion of attracted PC-MC pairs (9.8% ± 2.9) is greater than that of dispersed PC-MC pairs (4.6% ± 2.3; T = 1, n = 7, p =.035). The increased level of affiliation in PC samples was also selective to former opponents. Former opponents accounted for 7.4% (± 1.6) of affiliative contacts during PC samples whereas they accounted for 2.1% (± 1.0) of affiliative contacts during MC samples (T = 0, n = 7, p = 0.02).

In the Assamese macaque group, we collected 247 PC-MC pairs on 19 adults (mean = 13.0, SE = 1.0, range = 8–21). We previously reported the occurrence of reconciliation, selective attraction between former opponents, and the overall mean conciliatory tendency for the group of Assamese macaques (Cooper and Bernstein 2002). We calculate conciliatory tendencies for each sex class here for the first time. The mean conciliatory tendency is not significantly different between the rhesus and Assamese macaque groups (Table III). In fact, the conciliatory tendency for Assamese female-female dyads is low and in the wrong direction compared to rhesus females based on the phylogenetic hypothesis.
Table III

Reconciliation

 

Macaca assamensis

Macaca mulatta

U

p

N

Mean ± SE

N

Mean ± SE

Conciliatory tendency (%)

Adults

19

11.2 ± 2.9

7

6.6 ± 2.8

42.5

ns

Male-male

9

17.6 ± 2.2

2

0.0 ± 0.0

Female-female

10

10.7 ± 7.6

5

12.7 ± 6.4

20.0

ns

Male-female

17

5.9 ± 5.5

7

1.4 ± 1.4

53.0

ns

Dyads indicate the sex class of opponents, but do not indicate actors and recipients because we did not record the initiation of reconciliation. Cooper and Bernstein (2002) previously reported portions of the data on conciliatory tendency in Assamese macaques.

Approach

We calculated the rate of approaches and the mean proportion of approaches that resulted in positive and negative outcomes (Table IV). In the Assamese group we recorded 1090 approaches (mean = 38.9, SE = 5.3, range = 7–134), and in the rhesus group we recorded 136 approaches (mean = 17.0, SE = 3.0, range = 5–27). Assamese macaques had a significantly higher rate of approaches than that of rhesus macaques (U = 46.5, p = 0.024). The proportion of approaches resulting in positive and negative outcomes does not significantly differ between Assamese and rhesus macaques. The rate of approaches by males toward females does not significantly differ between species; however, the approaches were more likely to result in negative outcomes in Assamese macaques than in rhesus macaques (U = 1, p = 0.048). Females approached males at a significantly higher rate in Assamese macaques than in rhesus macaques (U = 8.5, p = 0.006), but the proportion of positive and negative approach outcomes is not significantly different. In female-female dyads, Assamese and rhesus macaques do not differ significantly in approach rate or approach outcome.
Table IV

Rate of approaches and approach outcomes

 

Macaca assamensis

Macaca mulatta

U

p

N

Mean ± SE

N

Mean ± SE

Rate of approaches

Adults

25

4.65 ± .70

8

2.09 ± .38

46.5

0.024

Male-male

11

2.00 ± .33

2

0.19 ± .06

Female-female

14

2.41 ± .46

6

1.94 ±.39

39.0

ns

Male-female

11

4.22 ± 1.10

2

1.63 ± .87

7.0

ns

Female-male

14

1.02 ± .15

6

0.27 ±.14

8.5

0.006

Negative outcomes (%)

Adults

25

19.7 ± 2.0

8

13.9 ± 5.5

59.5

ns

Male-male

11

15.2 ± 3.4

2

25.0 ± 25.0

Female-female

14

16.7 ± 4.2

6

15.4 ± 9.4

37.0

ns

Male-female

11

32.6 ± 5.4

2

10.9 ± 5.8

1.0

0.048

Female-male

14

10.1 ± 2.6

3

6.7 ± 6.7

17.5

ns

Positive outcomes (%)

Adults

25

24.3 ± 2.1

8

19.2 ± 5.1

76.0

ns

Male-male

11

25.3 ± 3.7

2

50.0 ± 50.0

Female-female

14

34.5 ± 5.4

6

22.0 ± 6.8

25.0

ns

Male-female

11

22.9 ± 3.0

2

13.4 ± 3.3

3.0

ns

Female-male

14

9.6 ± 2.7

3

22.2 ± 14.7

14.5

ns

Dyads indicate actors followed by recipients. In the rhesus group, the sample size for female-male approach outcomes is reduced because 3 females did not approach males.

Grooming

Previously we reported data on the distribution of grooming in the Assamese macaques (Cooper and Bernstein 2000); however, the values presented here differ because we restricted our analysis to adults. We calculated the rate of grooming between adults in Assamese and rhesus macaques and found that it did not significantly differ for adult, female-female, male-female, and female-male dyads (Table V). Because we collected many more grooming interactions between adults ad libitum than during focal samples, we combined data obtained ad libitum and focal sample data for analysis of the mean proportion of interactions. Also, mean proportions have the advantage of correcting for group size. In the Assamese group we recorded 3244 grooming interactions (mean = 129.8, SE = 24.0, range = 12–494) and in the rhesus group we recorded 741 grooming interactions (mean = 92.6, SE = 20.5, range = 34–207). Assamese and rhesus males do not differ significantly in the mean proportion of grooming they direct toward other adults (Table V). However, the mean proportion of grooming performed by rhesus females is greater than it is for Assamese females (U = 3.0, p = 0.001). Assamese females directed 79.5% (± 4.8) of their grooming toward other females, and rhesus females directed 83.5% (± 8.6) of their grooming toward other females (p = 0.46). Thus, despite the higher proportion of males in the group, Assamese females did not groom males proportionately more often than the rhesus females did.
Table V

Grooming

 

Macaca assamensis

Macaca mulatta

U

p

N

Mean ± SE

N

Mean ± SE

Rate of grooming

Adults

25

1.06 ± 0.18

8

0.84 ± 0.21

92.0

ns

Male-male

11

0.20 ± 0.05

2

0.0 ± 0.0

Female-female

14

1.10 ± 0.28

6

0.81 ± 0.29

37.5

ns

Male-female

11

0.63 ± 0.21

2

0.44 ± 0.06

10.5

ns

Female-male

14

0.14 ± 0.04

6

0.17 ± 0.11

38.0

ns

Proportion of total grooming (%)

Males

11

3.7 ± 1.4

2

4.6 ± 0.0

6.0

ns

Females

14

4.2 ± 0.8

6

15.1 ± 2.9

3.0

0.001

Shannon-Weaver heterogeneity index (H′)a

Adults

25

.391 ±.018

8

0.559 ± 0.061

43.0

0.017

Male-male

11

.390 ±.055

 

 

Female-female

14

.560 ±.034

6

0.792 ± 0.064

11.0

.011

Proportion of dyads grooming up dominance hierarchy (%)

Adults

25

37.7 ± 4.0*

8

49.8 ± 7.3

 

 

Male-male

11

55.1 ± 7.8

2

0.0 ± 0.0

 

 

Female-female

14

34.9 ± 4.6*

6

53.3 ± 6.7

 

 

Dyads indicate actors followed by recipients.

aWe corrected the Shannon-Weaver heterogeneity index (H′) for group size with the Buzas and Gibson’s (1969) evenness index. The H′ value for rhesus male-male dyads is unavailable because there was 1 dyad only.

*Significantly different from chance (sign tests, p < 0.01).

We examined each individual’s grooming distribution with the Shannon-Weaver heterogeneity index (H′) corrected for group size (Table V) (Buzas and Gibson 1969). Rhesus macaques distributed their grooming more evenly than Assamese macaques did (U = 43.0, p = 0.017). The effect was also evident when we restricted the analysis to female-female dyads (U = 11.0, p = 0.011).

We also analyzed the mean proportion of dyads that directed more grooming up the dominance hierarchy than down it. In the Assamese macaque group, dyads directed grooming down the dominance hierarchy more often than expected by chance. Though one would expect 50% of dyads to direct more of their grooming up the dominance hierarchy, only 37.7% of dyads directed more grooming up the dominance hierarchy (Table V). Assamese female-female dyads directed more grooming down the dominance hierarchy than expected by chance, whereas Assamese male-male dyads showed no significant bias in the direction of grooming (Table V). In the rhesus group, there is no significant bias in the direction of grooming for any of the dyad classes.

Male-Infant Contact

We observed 566 and 40 dyadic male-infant affiliative interactions in Assamese and rhesus macaques, respectively. Though the values are not comparable, because we collected data ad libitum on many more Assamese males and for a greater amount of time, dyadic male-infant affiliative interactions occurred regularly in both species. One of the most striking differences between the species was that triadic male-infant interactions were relatively common in the Assamese group (occurring 1.2 times per 12-h day), but did not occur in the rhesus group. Statistically, the proportion of male-infant affiliative interactions that included triadic interactions is greater in Assamese macaques (18.3% ± 5.3) than in rhesus macaques (0.0% ± 0.0; U = 1, p = 0.047).

Discussion

Female Dominance Style

Assamese females do not differ from rhesus females on several measures of dominance style, suggesting that Assamese macaques have a despotic dominance style. Further, our results support previous research showing that measures of dominance style covary as a set of traits (Castles et al. 1996). Also, because Assamese macaques belong to the sinica lineage, whose species tend to display tolerant dominance relationships, our results contradict a strong role for phylogenetic inertia in shaping dominance style in the sinica group. Because the extent to which our study group corresponds to other Assamese populations is unclear, one should treat data from a single group with appropriate caution.

Female-female dyads in both Assamese and rhesus macaques had low conciliatory tendencies (11% and 13%, respectively) and low proportions of counter aggression (6% and 3%, respectively). Reconciliation and counter aggression are common measures of dominance style and together the low levels suggest that Assamese and rhesus females have despotic social relationships. Our sample of agonistic behavior, including reconciliation, in rhesus macaques was small. However, our data are consistent with previous reports (de Waal and Luttrell 1989; Thierry 1985). Also, the proportion of aggressive interactions with counter aggression ranges from 0 to 12% for female-female dyads in despotic macaques (Berman et al. 2004; Castles et al. 1996; Petit et al. 1997; Thierry 1985), whereas it is often >50% for female-female dyads in species with a tolerant dominance style (Petit et al. 1997; Thierry 1985). Similarly, conciliatory tendencies range from 4% to 14% for females in despotic macaque species (Aureli et al. 1997; Berman et al. 2004; Call et al. 1999; Petit et al. 1997), whereas they range from 25% to 40% for females in tolerant macaque species (Abegg et al. 1996; Aureli et al. 1997; Call et al. 1999; Petit et al. 1997). Many aspects of macaque social behavior, including dominance style traits, are strongly influenced by matrilineal kinship. Because we had no information on kin relationships, we are unable to address the important affects of kinship on dominance style and our data are more appropriately compared to data from the literature that combine kin and nonkin. In the case of conciliatory tendency, interspecific comparisons are often made between nonkin females, and if we exclude kin from our data set, the resulting conciliatory tendencies for females would likely be even lower.

Assamese and rhesus females appeared equally despotic on other types of agonistic behavior. The directional inconsistency index for submissive behavior was low for female-female dyads in both the Assamese (1.4%) and rhesus (0) groups. The number of submissive interactions below the matrix diagonal is often <5% in despotic or nepotistic societies (Berman et al. 2004; Isbell and Young 2002), and thus both groups appear despotic on this measure also.

Female-female dyads in the 2 species are not significantly different in rates of aggression and the proportion of aggressive interactions involving bites, and the data are comparable to previous findings on other captive groups of rhesus macaques (Bernstein and Ehardt 1985; de Waal and Luttrell 1989; Thierry 1985). Likewise, there is no significant interspecific difference in the proportion of aggressive interactions with severe biting. Polyadic aggressive interactions occurred often in each group, suggesting that coalitions are an important aspect of female intragroup competition. Assamese females tended to aid other females more often than rhesus females did, possibly to avoid male aggression. In Assamese macaques, polyadic aggressive interactions were likely to involve bidirectional aggression among multiple males, whereas in rhesus macaques, females opportunistically aided males against lower-ranking females.

Assamese and rhesus macaques also showed a despotic pattern on several measures of affiliation. Female-female dyads in both Assamese and rhesus macaques showed a high proportion of approaches with negative outcomes (17% and 15%, respectively). The results for negative approach outcomes are consistent with the ones reported for rhesus macaques (de Waal and Luttrell 1989), and an established, cliquish group of pigtailed macaques (Castles et al. 1996). For female-female dyads the rate of grooming does not differ significantly between species, but rhesus females accounted for a greater proportion of grooming performed than Assamese females did. Contrary to our prediction, rhesus females distributed their grooming to other females more evenly than Assamese females did. Neither species groomed up the dominance hierarchy in a way that would be consistent with grooming for future favors (Seyfarth and Cheney 1984). In fact, Assamese females directed grooming down the hierarchy, which is consistent with our previous findings, a portion of which we used here (Cooper and Bernstein 2000). In sum, our results suggest that several types of agonistic and affiliative behavior are similar in Assamese and rhesus females, and are comparable to that reported for other despotic macaques (Berman et al. 2004; de Waal and Luttrell 1989; Petit et al. 1997; Thierry 1989). In addition, our results are consistent with the hypothesis that measures of dominance style covary as a set of traits (Castles et al. 1996; Thierry 2000).

Our results are in contrast to what we expected based on evidence of a strong correlation between phylogeny and macaque dominance style (Matsumura 1999; Thierry et al. 2000). Prior data suggest that despotic dominance styles are restricted mainly to the fascicularis lineage. However, the number of species outside the fascicularis lineage reported to show despotic dominance styles is growing, and the exceptions weaken a phylogenetic explanation for macaque dominance styles. Pigtailed macaques, of the silenus-sylvanus lineage, exhibit several characteristics associated with despotic relationships (Castles et al. 1996). Further, Tibetan (Macaca thibetana) and Assamese macaques are members of the sinica lineage, and both species appear to have a despotic dominance style, particularly among females (Berman et al. 2004). However, the significance of despotic dominance styles in the sinica lineage depends in part on the phylogenetic tree one adopts. In (Delson’s 1980) phylogenetic tree based on morphology, Assamese and Tibetan macaques are placed together as sister species with a recent common ancestor. Thus, a single reversal from a tolerant ancestral condition could explain the current dominance style of each species. In Hoelzer and Melnick’s (1996) tree based on mitochrondial DNA, Assamese macaques are divided into an eastern and western subspecies, which correspond to Macaca assamensis assamensis and M. a. pelops. Only the eastern subspecies is closely related to Tibetan macaques, and the western subspecies is probably a remnant population of bonnet macaques that subsequently hybridized with male Assamese macaques. Though our group of Assamese macaques was near the border between M. a. assamensis and M. a. pelops, their long tails suggest that they are M. a. pelops. Thus, per Hoelzer and Melnick’s (1996) scenario, a despotic dominance style for Macaca assamensis pelops might represent an additional reversal from precedent tolerant Macaca radiata.

If traits related to dominance style covary owing to evolutionary selective pressures, they should be characteristic of the species. Several measures of dominance style are based on proportions of behavior and are intended to remain independent of such factors as food availability, reproductive season, spatial density, and group size. In fact, several studies have shown that individuals in wild populations have patterns of postconflict behavior similar to those in captive groups (Aureli 1992; Castles and Whiten 1998; Kutsukake and Castles 2001; Matsumura 1996; Wittig and Boesch 2005). Further, Thierry (1985) and de Waal and Luttrell (1989) performed the original studies on indicators of dominance style on captive groups living under similar conditions, and they found consistent interspecific differences. The similar environments of our study groups should have removed some sources of variation and made the similarities in agonistic and affiliative behavior all the more striking. Still, a few important caveats remain. Data collection occurred partly during the mating season in the Assamese group, and partly during the birth season in the rhesus group. Some of the interspecific differences in male-female behavior are possibly due to the reproductive season, but the degree of affiliation between males and females was greater in the rhesus group during the nonbreeding season than it was in the Assamese group during the breeding season. Also, the Assamese group was much larger than the rhesus group. Rates of social behavior vary with group size (Koenig et al. 2004; Sinha et al. 2005), but proportional measures such as reconciliation should be less affected by group size (Castles et al. 1996; de Waal and Luttrell 1989). Similarly, we were unable to address how differences in group size might interact with kinship to produce different distributions of kin and subsequently affect measures of dominance style.

The possibility remains that some of the similarities in agonistic and affiliative behavior we observed between Assamese and rhesus macaques were due to the subjects living in similar habitats, and habitat-dependent traits need to be reevaluated for their usefulness as measures of dominance style. Provisioned environments that generate intense feeding competition may make groups appear more aggressive and more despotic during feeding. If the goal is to understand the evolution of primate social behavior, we must recognize intraspecific variation in agonistic and affiliative behavior so that we can better understand species-specific indicators of dominance style. For instance, both study groups were residents of temples and had frequent contact with people, and several individuals were bitten at each study site, giving local people the impression that monkeys are mean. However, how monkeys interact with people is not necessarily a good representation of how they interact with one another and does not indicate dominance style. In both study groups, the rate of aggression and proportion of bites were similar to the ones in a variety of Old World monkeys in captive groups (Bernstein et al. 1983), suggesting that the rate of aggression and the proportion of bites are probably not good indications of a species-specific dominance style. Likewise, exceptions to Kawamura’s (1965) principles for the inheritance of dominance rank may be associated with the age-structure of the group rather than dominance style (Chikazawa et al. 1979; Datta and Beauchamp 1991; Nakamichi et al. 1995). Also, affiliation between adult males, including triadic male-infant interactions, may not reflect dominance style as defined by female social relationships.

Male Social Relationships

Researchers have historically determined dominance style by the social behavior of the philopatric sex, which in macaques is female. However, with more data accumulating, sexual differences in dominance style are now recognized (Berman et al. 2004; Cooper and Bernstein 2002; Preuschoft et al. 1998; Wittig and Boesch 2003;). In our study, Assamese males showed tolerant levels of reconciliation and counter aggression, suggesting that they do not easily fit with the despotic dominance style of Assamese females.

Conciliatory tendencies for other macaque males that have strong affiliative relationships, such as Tibetan and bonnet macaques, are 19.7% and 30.2%, respectively (Berman et al. 2004; Cooper et al. 2007). The Assamese males of our study had a conciliatory tendency of 17.6%, which is within a tolerant range. Rhesus and Japanese macaque males generally have weak affiliative relationships; however, substantial intraspecific variation occurs in conciliatory tendencies of Japanese macaques (9.7–31%: Majolo et al. 2005; Petit et al. 1997; Schino et al. 1998). Thus, one should take values from single groups with caution. The proportion of counter aggression by Assamese males (25.4%) is greater than those of rhesus, long-tailed (Macaca fascicularis), and Tibetan macaque males (Berman et al. 2004; Thierry 1985), though it is not nearly as great as among Tonkean males (M. tonkeana: Thierry 1985).

Dyadic male-infant affiliative contact occurred in both Assamese and rhesus macaques. However, in Assamese macaques, male-infant contact included triadic male-infant interactions which resemble interactions in Barbary (Macaca sylvanus: Taub 1980) and Tibetan macaques (Ogawa 1995). In contrast, dyadic male-infant contact in rhesus macaques involved mainly retrieval from threatening situations with humans, and did not involve triadic male-infant interactions. Further, Assamese males approached and groomed each other, and often aided one another in fights. Rhesus males seemed to express such affiliative and coalitionary relationships less often, though it is difficult to reach any firm conclusion owing to the smaller rhesus male sample.

There are several possible explanations for variation in male dominance style. In Barbary macaques, contests between males for a peanut reward often end in a stalemate whereas females tend to assert their dominance over other females (Preuschoft and Paul 2000). Male Barbary macaques are among the largest macaque males (Smith and Jungers 1997), and their stalemate may indicate a reluctance to engage in a potentially dangerous fight over a small incentive. Thus, when males are large and possess dangerous weaponry, such as large canine teeth, individuals may not express dominance relationships because the risk of escalation is too great. The potential costs of fighting may partly explain a tolerant dominance style among male Assamese macaques, which are relatively large versus other macaque males (Smith and Jungers 1997), and they have relatively large canine teeth (Hill and Bernstein 1969; Thoren et al. 2006). Another important factor is that as male aggression escalates it is likely to become a polyadic interaction in Assamese macaques. Thus, in Assamese males, willingness to assert dominance or to contest existing relationships may depend on the availability of coalition partners.

Another important factor affecting male social relationships is the number of potential partners. In the present study, if one considers only fully adult individuals, the sex ratio is less skewed in the Assamese macaque group than in the rhesus group. In rhesus populations extragroup males are solitary, peripheral males, whereas we saw no solitary male during our preliminary surveys of Assamese macaque distribution. It seems likely that having more potential partners allows for, or may even require, the expression of tolerant dominance relationships and strong male bonding. Berman et al. (2006) suggested that the nearly even sex ratio, rather than the high degree of sexual dimorphism, may account for the increased male-male tolerance in Tibetan macaques. However, it is difficult to separate the influence of body size and sex ratio in Assamese, Tibetan, and Barbary macaques because all 3 species have large males (Smith and Jungers 1997) and nearly even sex ratios (Fooden 1986; Menard and Vallet 1996; Zhao and Deng 1988). However, bonnet macaques provide a potentially more informative example. Male bonnet macaques have a high conciliatory tendency (Cooper et al. 2007), frequently groom and aid one another (Silk 1994), and engage in triadic male-infant interactions (Silk and Samuels 1984). Thus, they appear to have strong bonds and a tolerant dominance style. Interestingly, bonnet males are rather small compared to other macaque males (Smith and Jungers 1997), and they live together with numerous males per group (Singh et al. 1984). Consequently, their tolerant dominance relationships may be unrelated to the high costs of dyadic aggression associated with large body size. In sum, our results for male Assamese macaques are consistent with an association between numerous adult males per group and a tolerant male dominance style. Further, male social relationships appear to be independent of female social relationships and the resulting species-specific dominance style.

Acknowledgments

A grant from the National Geographic Society (grant 5862–97) funded the research, which we conducted in cooperation with the Indo-US Primate Project, principal investigators S. M. Mohnot and Charles Southwick. We thank Arun Srivastava and Prabal Sarkar for logistical support in the field, and Mohibul Haque for help in data collection. The University of Georgia Animal Care and Use Committee approved all procedures, and the study complied with Indian law.

Copyright information

© Springer Science+Business Media, LLC 2008