intergroup interactions in tibetan macaques at mt. emei, china

12
Intergroup Interactions in Tibetan Macaques at Mt. Emei, China QI-KUN ZHAO Kunming Institute of Zoology, Chinese Academy of Sciences, Kunming, Yunnan 650224, People’s Republic of China KEY WORDS close encounter; fission; food competition; long- distance interaction; Macaca thibetana; sexual attractiveness; sexual selection ABSTRACT Data on intergroup-interactions (I-I) were collected in 5 seasonally provisioned groups (A, B, D, D 1 , and E) of Tibetan macaques (Macaca Thibetana) at Mt. Emei in three 70-day periods between 1991 April–June (P1), September–November (P2), December–1992 February (P3). The I-I were categorized as forewarning made by high-ranking males (including Branch Shaking and/or Loud Calls), long-distance interactions in space (specified by changes in their foraging movements), and close encoun- ters (with Affinitive Behavior, Male’s Herding Female, Sexual Interaction, Severe Conflict, Adult Male-male Conflict, Opportunistic Advance and Re- treat, etc. performed by different age-sex classes). From periods P1 to P3, the I-I rate decreased with reduction in population density as a positive correlate of food clumpedness or the number of potential feeders along a pedestrian trail. On the other hand, from the birth season (BS, represented by P1 and P3) to the mating season (MS, represented by P2) the dominance relation between groups, which produced a winner and a loser in the encounters, became obscure; the proportion of close encounters in the I-I increased; the asymme- try (local groups over intruders) of forewarning signals disappeared; the rate of branch shaking decreased; and sometimes intergroup cohesion appeared. Considering that sexual interactions also occurred between the encountering groups, above changes in intergroup behaviors may be explained with a model of the way in which the competition for food (exclusion) and the sexual attractiveness between opposite sexes were in a dynamic equilibrium among the groups, with the former outweighing the latter in the BS, and conversely in the MS. Females made 93% of severe conflicts, which occurred in 18% of close encounters. Groups fissioned in the recent past shared the same home range, and showed the highest hostility to each other by females. In conspicuous contrast with females’ great interest in intergroup food/range competition, adult male-male conflicts that were normally without body contact occurred in 66% of close encounters; high-ranking male herding of females, which is typical in baboons, appeared in 83% of close encounters, and showed no changes with season and sexual weight-dimorphism; peripheral juvenile and subadult males were the main performers of the affinitive behaviors, opportunistic advance and retreat, and guarding at the border. In brief, all males appeared to ‘‘sit on the fence’’ at the border, likely holding out hope of gaining the favor of females both within and outside the group. Thus, Contract grant sponsor: National Geographic Society; contract grant number 402289; Contract grant sponsor: National Natural Science Foundation of China; contract grant number 38970147; Contract grant sponsor: Chinese Academy of Sciences; contract grant number KZ 951-A1-105. Received 25 August 1993; accepted 10 August 1997. AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 104:459–470 (1997) r 1997 WILEY-LISS, INC.

Upload: qi-kun

Post on 06-Jun-2016

213 views

Category:

Documents


1 download

TRANSCRIPT

Intergroup Interactions in Tibetan Macaques at Mt. Emei, ChinaQI-KUN ZHAOKunming Institute of Zoology, Chinese Academy of Sciences, Kunming,Yunnan 650224, People’s Republic of China

KEY WORDS close encounter; fission; food competition; long-distance interaction; Macaca thibetana; sexual attractiveness; sexualselection

ABSTRACT Data on intergroup-interactions (I-I) were collected in 5seasonally provisioned groups (A, B, D, D1, and E) of Tibetan macaques(Macaca Thibetana) at Mt. Emei in three 70-day periods between 1991April–June (P1), September–November (P2), December–1992 February (P3).The I-I were categorized as forewarning made by high-ranking males(including Branch Shaking and/or Loud Calls), long-distance interactions inspace (specified by changes in their foraging movements), and close encoun-ters (with Affinitive Behavior, Male’s Herding Female, Sexual Interaction,Severe Conflict, Adult Male-male Conflict, Opportunistic Advance and Re-treat, etc. performed by different age-sex classes). From periods P1 to P3, theI-I rate decreased with reduction in population density as a positive correlateof food clumpedness or the number of potential feeders along a pedestriantrail. On the other hand, from the birth season (BS, represented by P1 and P3)to the mating season (MS, represented by P2) the dominance relation betweengroups, which produced a winner and a loser in the encounters, becameobscure; the proportion of close encounters in the I-I increased; the asymme-try (local groups over intruders) of forewarning signals disappeared; the rateof branch shaking decreased; and sometimes intergroup cohesion appeared.Considering that sexual interactions also occurred between the encounteringgroups, above changes in intergroup behaviors may be explained with a modelof the way in which the competition for food (exclusion) and the sexualattractiveness between opposite sexes were in a dynamic equilibrium amongthe groups, with the former outweighing the latter in the BS, and converselyin the MS. Females made 93% of severe conflicts, which occurred in 18% ofclose encounters. Groups fissioned in the recent past shared the same homerange, and showed the highest hostility to each other by females. Inconspicuous contrast with females’ great interest in intergroup food/rangecompetition, adult male-male conflicts that were normally without bodycontact occurred in 66% of close encounters; high-ranking male herding offemales, which is typical in baboons, appeared in 83% of close encounters, andshowed no changes with season and sexual weight-dimorphism; peripheraljuvenile and subadult males were the main performers of the affinitivebehaviors, opportunistic advance and retreat, and guarding at the border. Inbrief, all males appeared to ‘‘sit on the fence’’ at the border, likely holding outhope of gaining the favor of females both within and outside the group. Thus,

Contract grant sponsor: National Geographic Society; contract grant number 402289; Contract grant sponsor: National NaturalScience Foundation of China; contract grant number 38970147; Contract grant sponsor: Chinese Academy of Sciences; contract grantnumber KZ 951-A1-105.

Received 25 August 1993; accepted 10 August 1997.

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 104:459–470 (1997)

r 1997 WILEY-LISS, INC.

females and males attempted to maximize reproductive values in differentways, just as expected by Darwin–Trivers’ theory of sexual selection. Inaddition, group fission was observed in the largest and highest-ranking groupfor two times (both in the MS) when its size increased to a certain level, andthe mother group kept their dominant position in size and rank among thegroups that might encounter, suggesting that fission takes a way of discardingthe ‘‘superfluous part’’ in order to balance the cost of competition for food andmates within a group, and the benefit of cooperation to access the resourcesfor animals in the mother group. Am J Phys Anthropol 104:459–470, 1997.r 1997 Wiley-Liss, Inc.

Female residence and systematic move-ment of males among groups occur in ma-caques and many other primate species. Thesocial unit of such species is typically ‘‘fe-male-bonded,’’ i.e., composed almost entirelyof female kin and immigrant males (Wrang-ham, 1980). The social group of Tibetanmacaques (Macaca thibetana) belongs tothis category (Zhao, 1994a).

Group living in primates has been pro-posed to be selected to enhance the successof individuals in defense of limited resourcessuch as food, water, and sleeping sites, thusthe intergroup relationship is essentiallycompetitive (Wrangham, 1980). However, ifthe diameter of the home range of a primategroup is larger than the average distancecovered daily, the group is likely to share itsrange, at least partly, with other groups(Mitani and Rodman, 1979). When home-range overlap is great, as it typically is inmacaques, aggressive encounters occur atlow rates, and those that do occur usuallyconcern access to a clumped resource, suchas that at human provisioning sites. Awayfrom such resources, macaque encountersgenerally involve passive avoidance (South-wick et al., 1965; Kawanaka, 1973; Lind-burg 1977).

Intergroup competition for resources isoften mediated by the relative dominanceranks of the groups involved. Dominanceamong macaque groups is usually a functionof group size, or the number of adult females(Wrangham, 1980). Occasionally, however, itmay depend on more subtle factors, such aspast relations among the male members ofdifferent groups in rhesus macaques (Gabow,1972).

In most mammalian species, females haveto channel more energy to offspring than do

males, and their reproductive success ap-pears limited primarily by food acquisitionrelated to the foraging range. In contrast,male reproductive success is limited mainlyby the availability of females (Darwin, 1871;Trivers, 1972). This is the essence of sexualselection theory. Accordingly, such a sexdifference is expected to be embodied in theperformances of females and males in inter-group interactions. That is, a) females shouldtend to engage in intergroup food/rangecompetition more actively and seriously thanmales, and b) males should try their best togain the favor of females both within, andoutside the group for additional mates imme-diately or in consideration of smoothinglater transfer.

In studies on primate intergroup interac-tions, the aspect of competition for food/range has been emphasized (see Cheney,1987), but little attention was paid to theopposite—the force contended against theexclusion—though sexual attraction to unfa-miliar opposite sexes is commonly observedin non-human primates (Enomoto, 1974;Sugiyama, 1976; Cheney and Seyfarth, 1977;Brereton, 1981; Hausfater, 1972; Zhao, 1993)and in humans (Shepher, 1971; Wolf andHuang, 1980), and male reproductive strate-gies have recently been suggested to governintergroup relationships in baboons (Cowl-ishaw, 1995).

At the Mt. Emei site, the Tibetan macaquegroups are living under an unintended natu-ral (unplanned) experimental condition withexaggerated seasonal changes in food avail-ability and clumpedness due to provision-ing. This, together with the directional move-ment of potential feeders, to which themacaque groups response in a way of mov-ing in the opposite direction (Zhao, 1994c),

460 Q.-K. ZHAO

and the involvement of as many as 5 groupsafford a rare opportunity to inspect thecurrent selective pressures on the socialgroup of Tibetan macaques, a previouslylittle known species. This study was plannedto approach this important but poorly docu-mented topic through analyses of a) therelation of intergroup interaction rates tofood clumpedness, b) the behavior of malesand females during intergroup interactions,and c) demographic data in relation to groupfission and the interactions between fis-sioned groups.

METHODS

The study site is mainly between 1,500–2,400 m on the northeast slope of Mt. Emei(29°308N and 103°198 for the summit) risingfrom 500 m to 3,099 m a.s.l. The foresthabitat of study groups is usually snow-blanketed for 3.5 months per year, and themean temperature in January is 21.3°C at2,070 m (Zhao, et al., 1989).

The macaques that range in or near thetourist and Buddhist center, depend on bothfood handouts and natural foods—foliage inspring and summer, and bamboo shouts andfruits in autumn—in the tourist or warmseasons, and otherwise, rely basically onfoliage (Zhao et al., 1991). Seasonal provi-sioning from tourists and Buddhists hasconsiderably exaggerated the periodical fluc-tuation of food availability and clumped-ness, and made the macaque groups spendmore and more time foraging along the trail.Because visitors enter the trail from eitherthe upper (2,400 m) or lower (750 m) end, 10groups ranged along the trail dependingtheir intergroup rank in access to the feed-ers, and categorized as high, middle, and lowgroups (Zhao, 1994d). The study groups (A,B, D, and D1) were those resident in the highpart of the site, dominant over the middlegroups, tending to forage daily by competi-tively moving upwards toward the descend-ing trailgoers (Zhao, 1994c). Sometimesgroup E, one of the low groups (also domi-nant over the middle groups), moved throughthe middle, and had interactions with thegroups B and D1.

Quantitative data on intergroup interac-tions were collected between 0800 h–1700 heach day in three successive 70-day periodsin the spring and summer (P1; April 15–

June 23, 1991), autumn (P2; September1–November 9, 1991), and winter (P3, De-cember 1–February 8, 1991/2). In the threeperiods, groups B, D, and D1 were radio-tracked. Over the seasons, the potentialhuman feeders were present in thousandseach day in spring and summer, but only inhundreds daily in autumn, and tens daily inwinter.

P3 and P1 can be used to represent thebirth season (Zhao, 1994d), in which themacaques are sexually inactive. Within thebirth season, the availability of human-provided food along the trail reaches itsannual peak between May and August, butfalls in the annual valley in winter. P2 fallsin the middle mating season, in which sexualactivities both within and between groupsreaches the peak, but the availability ofhuman-provided food falls between P1 andPs.

Data on intergroup encounters includethose observed between/among five groups(A, B, D, D1, and E) when they entered theregular observation areas (trail sections c, d,

Fig. 1. Trail sections used by the study groupsranging at the upper part of the population range. Keys:used (thick lines a–i) and unused (dot lines) trail sec-tions, temples visited by the groups (T1–4), streams (thinline), and elevation above sea level in meters (four-digitnumber).

461INTERGROUP INTERACTIONS IN TIBETAN MACAQUES

e, f, g in Fig. 1), the groups’ activity center.Survey routes were usually taken accordingto the following schedule: if any of thegroups B, D1 and E appeared at sections e, f,and g near temple T4, my camp, I stayedthere with one of the groups (but not morethan three successive days for one group),then changed to another group. After work-ing in this area, I moved to follow the groupsin sections c and d (A, B, D and D1) in thesame way. Other sections (a, b, h, and i) werealso surveyed, but not regularly. With theaid of the radio-tracking system, some of thelong-distance interactions among groups B,D, and D1 were determined by checkingtheir daily ranging routes. Following thesurvey program, the relative rate of closeencounters and long-distance interactions(see below) should be comparable across thesampling periods.

The occurrence of a particular behaviorwas scored as 1 when it occurred at leastonce in the course of an interaction, andotherwise as 0. Because not all encounterswere fully observed, the one-zero scores prob-ably described the occurrence of these behav-iors more accurately than the exact frequen-cies recorded in the course of imperfectlyobserved encounters. On the other hand, theobservation duration for close encountersshowed no difference between the birth andmating seasons (2.4 vs. 2.9 h, t 5 20.75,P . 0.05), and the occurrence rate of behav-iors is likely to be comparable between theseasons.

Taking advantage of the fact that thegroups regularly moved up along the trail tomeet the tourists going down and carryingfood (Zhao, 1994c), the outcome of inter-group food competition was simply inferredfrom the groups’ foraging movements. Thatis, the dominant group controlled the trail orthe higher section, and the loser was drivenaway from the trail or remained in the lowerpart.

Because I did not always arrive at thebeginning of the interactions, and the en-countering groups sometimes soon disap-peared in the forest/cliff, sampling sessionsused in data processing include relativelycomplete observations only for either warn-ing phases or later encounters (see below).In addition to the quantitative data on the

interactions, some conspicuous events oc-curred outside the sampling periods are alsointegrated in this analysis for a full view ofthe encounters.

DEFINITIONSForewarning

These display patterns consisted of branchshaking and/or special long-loud calls (spec-trographs in prep), often directed by alphamales to an opposing group at a distance ofhundreds or tens of meters. Between twogroups in interaction, forewarning signalsmight be uttered by one group or exchangedby the two. Forewarning might or might notappear at the beginning of interactions de-scribed below.

Long-distance interactions

In these interactions, one group’s foragingmovement was influenced by another groupat a distance of hundred or tens of meters. Inthis category, a group might stop below atrail section occupied by a higher-rankinggroup, sometimes for days; go back downalong the trail or through the forest when ahigher-ranking group was detected at ahigher position, or moving in the oppositedirection; enter a trail area after the higher-ranking group fully left (for example, 30 minlater), leave a contested area, to which adominant group was moving, or retreat to anarea that was exclusively used by the group.

Close encounters

Situations in which two or three groupswere in physical contact or separated by adistance of less than 10 m, either seenclearly at the trail or judged to be in directcontact by the observer’s hearing specialfighting calls. This pattern might last for upto 5 h.

Border

In close encounters, an imaginary linemight be put between animals from differ-ent groups. Because home ranges over-lapped extensively, the border was a movingor soft one (see Opportunistic advance andretreat) in any part of the range. It was alsothe place for the observer to collect data onthe encounter.

462 Q.-K. ZHAO

Affinitive behaviors

At the border, animals from differentgroups made present, mount, mouth-mouthcontact, penis look (in which a male took thepresenter’s penis and looked at it for a shorttime), and contact threat [in which the domi-nant grasped at another animal that ac-tively presented/exposed the body part withsubmissive face expression, and lightly bitits front part of body, or grasped at head hairand then made an attack face (Bertrand,1969) at the latter]. In intragroup interac-tions, these affinitive behaviors, especiallythe contact threat, usually resulted in adecline of social tension between the ani-mals, reflected for example in the tolerationof a subordinate in proximity or feedingnearby.

Herding

Males tried to counter-chase or drive theirown females in order to keep them awayfrom the border. Females usually ran back-wards and waited to return to the borderagain when the herder was out of view, butthe situation was different in severe con-flicts (defined below).

Sexual interaction

Sexual behaviors occurred between ani-mals from different groups, consisting ofsexual initiation (bared teeth chatter ofmales, lifting female’s hip, female’s present-ing), and mating (mount, intromission, andpelvis thrust and/or ejaculation). Becausethis species is a multiple-mount-to-ejacula-tion species, mounting might or might notresult in ejaculation (Zhao, 1993). This typeof interaction was observed only in the mat-ing season.

Severe conflict

If males failed to herd females, and mostof a group’s members took part in an encoun-ter seriously, this category was scored. Suchconflicts, which might result in woundingand/or more severe consequences (describedbelow), were always accompanied by high-pitched calls uttered mainly by female fight-ers of both sides, and lasting up to 2.5 h.

Adult male–male conflict

A burst of violent male-male chasing, ac-companying a string of short-loud ‘‘co-co-co. . .’’ calls (spectrographs in prep) uttered bythe attacker. Usually, only two or threemales (excluding the very low-ranking ones)from different groups were involved in suchlocal fights, which never resulted in wound-ing.

Opportunistic advance and retreat

Silent actions at the border, in which anoutnumbered group retreated when an op-posing group was strengthened by one ortwo newcomers. The movement was re-versed if the retreating group received rein-forcements that reversed the balance. Theseesaw battle was mainly made by juvenileand subadult males, and might be repeatedmany times in a regular close encounter.

Intergroup dominance

If a close encounter resulted in a group’sretaining possession of, or displacing an-other group at, a favorite trail section, thekeeper or displacer was defined as the win-ner, the other was the loser. Otherwise, theclose encounter resulted in uncertainty. Fromthe tally of wins and losses, the intergroupdominance hierarchy could be determined.

Intruder and resident

If a group came into a close encounterwith another group that appeared earlier ata trail section, the later comer and the earlyone were defined as the intruder and resi-dent respectively.

RESULTS

Two groups came into interactions witheach other in a distance of hundreds or tensof meters when the foraging movement ofany one group appeared to be influenced byanother. A forewarning phase sometimesappeared at the beginning, and the groupsmight stay in long-distance interactions forup to days, or proceed to a close encounter, inwhich behaviors such as those listed aboveoccurred. At the end, a close encounter mightresult in a winner and a loser, or in uncer-tainty.

463INTERGROUP INTERACTIONS IN TIBETAN MACAQUES

Behavioral changes with season

The rate of long-distance interactions(LDI) plus close encounters (CE) (Table 1)was significantly different in spring-sum-mer (1.09/day), autumn (0.84), and winter(0.23) (F-ratio 5 31.04; df 5 2, 207; P ,0.001). In fact, the rate progressively de-creased with reduction of food clumpedness(described with the number of potentialfeeders) over the seasons.

On the other hand, the ratio of CE/LDIchanged from 19/57 in P1 (spring-summer),to 32/27 in P2 (autumn), to 7/9 in P3 (win-ter). The ratio of CE/LDI was 13/33 or 26/66(combination of P1 and P3) in the birthseason and 32/27 in the mating season, andthe change between the two seasons wassignificant (Proportion Test for CE/(CE 1 LDI): z 5 23.20, P , 0.001).

Branch shaking appeared more often inthe birth season than in the mating season(Table 2a). The resident groups tended toproduce forewarning signals (branch shak-ing and/or long-loud calls) more often thanthe intruder in the birth season, but not inthe mating season though the asymmetrywas retained in the year. In addition, theasymmetry became stronger if both branch

shaking and loud-calls were considered(Table 2b). Interestingly, 90% of 21 observedbouts of branch-shaking were made by alpha-males, 1/5 of the actions with long-loud calls.Of 7 forewarning phases with calls, 5 did notdevelop into close encounters.

Sexual interactions between animals fromdifferent groups were observed in 6 of 14close encounters in the mating season, andnone in the birth season. The sexual behav-iors included 9 matings (1 ejaculation) and16 initiations (8 bared-teeth chatters and 2lifting female’s hip of males, and 6 present-ing actions of females). Of the mounts, onlyone was made by an alpha male, and theejaculation was made by a subadult male(Table 2c).

A dominance hierarchy was observed in16 close encounters (Table 3) in four groupsin the birth season. Obviously, the rankorder for groups that met each other wasD . A . B . D1. However, the clear domi-nance relations became obscure in the mat-ing season (outcome in Table 2c).

Taking advantage of the changes ap-peared in the mating season, group D1 peace-fully made its first range extension intogroup B’s range three years after its fissionfrom group D in September, 1991.1

It is of interest to note how a rank-reversal between groups A and B occurred.In the 1986 birth season, group B wasdominant over group A under the conditionthat there were 8 females (F), 4 males (M), 2subadult males (SA) in A [group size (gs) 524], and 10 F and 5 M in B (gs 5 30, Zhaoand Deng, 1988). At the end of 1987, all 5adult males of group B were former mem-bers of group A. Of the five, two young adultmales Ne and Ni natally transferred in the1987 mating season, and two middle-agedmales Ha and Ey transferred after losingtheir alpha positions in 1985 and 1987 re-spectively. After the replacement of males ingroup B, Ni became the alpha, Ne also rose

1Mediated by one of two transferring males from D1 to B, ahigher-ranking group, D1 followed B moving around T4 (Fig. 1),where food handouts were relatively rich and D1 had neverarrived, for 3 days on 6–8th without conflict. In the period, D1always followed B in a distance less than 10 m, foraging attrail-sections e, g, and f or in the forest; sexually active animalsfrom different groups kept trying to have sexual interactionsduring the day. D1 even kept following B to its sleeping site in lateafternoon. After the exploration, D1 began to range in the areaaround T4 regularly.

TABLE 1. Distribution of long-distance interactions(LDI) and close encounters (CE) observed in five groupsof seasonally provisioned Tibetan macaques in three 70

day periods in spring–summer, autumn, and winterrespectively

Season

LDI CE

A B D D1 A B D D1

Spring & summer(1991)

AB 2 1D 13 4D1 2 11 19 6 3E 3 7 3 2Total (57) (19)

Autumn (1991)AB 3 1D 2 6 1D1 1 3 15 9 12E 3 2 1Total (27) (32)

Winter (1991/2)ABD 3 1D1 3 1 5 3ETotal (9) (7)

464 Q.-K. ZHAO

in rank (Zhao, 1994a), and A became domi-nant over B although the demographic vari-ables remained as B . A (7F and 4M in A; 8Fand 6M in B; size A/size B 5 29/36 in themiddle of the 1987 mating season; Zhao1993). The dominance of A over B continuedto the end of 1991 even though a newimmigrant from group D took the alphaposition of B when Ni and Ne died ofdropping down a large cliff accidentally andEy disappeared in the icy December of1990, and the size ratio further decreased to17/27.

Behaviors not changing with season

Males herding females at the border oc-curred in 83% of close encounters, butshowed no difference between the BS andthe MS (Table 2c, Proportion Test, z 5 20.41,P . 0.05). Ninety-one of 133 herding actionswere made by alpha males towards theiradult females.

Severe conflicts, which mainly involvedfemales, occurred in 18% of close encoun-ters. On the other hand, chasing betweenadult males from different groups occurredin 66% of close encounters (Table 2c). Inaddition, no intergroup male-female aggres-sion were observed in this study. Different

roles of females and males in the fight can beclearly seen.2

The peripheral males (mainly the juve-nile, subadult) were the most active ele-ments in affinitive performances, opportunis-tic advance and retreat (Table 2c), and inguarding at the border in normal close en-counters. In one case, a subadult male wasforced to be a guard at the border.3

Group fission and interactions betweenfissioned groups

Group fission was noted in the largest andmost dominant group (D) twice during thestudy period between 1986–1992. The fis-sions occurred when group size reached 79in 1988, and 61 in 1991 respectively. Bothoccurred in the mating season. In addition,the majority or the mother group (67 and 43respectively) kept its dominant position inboth size and rank among the groups—the

2On April 15, 1989, at 1300 h, 2 dominant adult females (withan infant) and a subadult male from group D1 faced, at a distanceof about 5 m, 2 mother-infant pairs and another female fromgroup D, about 7 m above section d (Fig. 1). Later, the alpha andbeta males of D1 arrived, and began to herd their females butfailed to drive them away from the contact zone. The femaleskept brandishing or moving one hand up and down in thedirection of opposing females at a distance less than 1 m, utteringa lot of high-pitched calls. The females’ fight and D1’s maleherding dislodged a lot of soil and stones from the slope. At 1418h, 3 more females from D joined the fight, and 6 females with 3infants formed a small wall at a higher position. Ten minuteslater, when two adult males appeared behind the females of D, ageneral attack from females of D began, and the border soonmoved down to the deep valley where the noisy calls lasted for 30min. The severest intergroup conflict resulted in (in addition toabout 120 kg of fallen soil and stones remaining on the trail) aninfant’s disappearance and the beta male’s transferring fromgroup D1. Interestingly, a low-ranking female and a subordinatemale from group D1 were not involved in the fight.

3On May 5th, 1986, 1245 h after two bursts of male-maleconflicts in the forest, the alpha male of group F forced a subadultmale to go backwards to the ‘border’ for about 15 m. Whenmoving backwards, the subadult faced the driver and grimacedrepeatedly. When the subadult reached the border and climbedinto a tree, the alpha male retired to a point about 25 m awayfrom the border. The subadult kept watching and sometimesmade lunge threats to his opponent for 30 min. At 1345 h, onemore male-male conflict occurred in the forest, but the guard atthe border seemed to be oblivious of the fight until C disappearedat 1400 h.

TABLE 2b. Seasonal changes in asymmetry offorewarming behaviors presented in Table 2a

(proportion test for difference between local groups andintruders, z-values and probability)1,2

Season(S.)

Branch-shaking

(BrS)Long-loudcall (LC) BrS & LC

Birth S. 2.74** 1.79* 4.03***Mating S. 0.94 0.00 0.83Both S. 2.74** 1.39 (NS) 4.14***1 The score in ‘Both’ column of Table 2a is added to scores for theresident and the intruder.2 * P , 0.05; ** P , 0.01; *** P , 0.001.

TABLE 2a. Seasonal changes in scores of forewarning behaviors in intergroup interactions in five groups of Tibetanmacaques at Mt. Emei 1

Season (S.)

Branch shaking Long-loud call

Resident Intruder Both Total Resident Intruder Both Total

Birth S. (n 5 18) 10 2 1 13 4 0 1 5Mating S. (n 5 12) 3 1 1 5 0 0 2 2Aver. rate of two S. (%) 40 10 7 57 11 0 11 22Difference between S.2 (z) 1.7** NS1 1 was scored when a behavior occurred in sampling period, otherwise, 0.2 Proportion test: ** P , 0.01; NS 5 P . 0.05.

465INTERGROUP INTERACTIONS IN TIBETAN MACAQUES

size of the second largest group (B) in thearea was 40 and 27 respectively, and that ofthe second highest-ranking group (A) was 33and 17.

Daughter group 1 (D1) was composed of 1middle-aged M, 3 F, and 8 immatures; anddaughter group 2 (D2) was of 1 young M, 1SA, 8 F, and 8 immatures.

Within 2–3 years after group fission, thesmaller daughter groups shared the originalrange with the larger one, but often took thelower/poorer trail sections, some of whichwere abandoned by the group before fission(Fig. 1 and Table 4). For example, group D1ranged usually at either section d or thelower g (using the forest between the twosections as passageway) when D was at c,otherwise, it would soon move to the bestfeeding site (c). Then the minority D2 re-peated the same ranging pattern with D1: itregularly ranged at or between sections cand i depending on the majority D’s pres-ence or absence at section c/d/h—if D ar-rived at or approached c/d/h, D2 usuallytook a long trip through the north-eastforest to section i, otherwise D2 soon re-turned to c. Note in particular that both gand i were given up by the un-fissionedgroup D, and the distance between i andc/d/h was far beyond the range of humansenses.

In competition for limited resources withinthe range, females from one part showed anextremely high hostility to those from theother. Of a few severe female-female con-flicts noted in the study periods since 1986,the most acute fights were observed betweenthe majority (D) and the minorities (D1 andD2), 6 and 2 months after the fissions respec-tively.

DISCUSSION

The rate of intergroup interactions de-creased from 1.09/day in spring-summer, to0.84 in autumn, and to 0.23 in winter whenthe population density changed from 44/km2

to 32 to 20 respectively (Zhao, unpublisheddata), or when the number of potentialfeeders decreased from thousands per day tohundreds to tens. The relation between popu-lation densities and interaction rates is ex-pected by Waser’s (1976) ‘gas’ model.

TA

BL

E2c

.S

easo

nal

chan

ges

inoc

curr

ence

sof

beh

avio

rsin

clos

een

cou

nte

rs1

Sea

son

(S.)

Adu

ltfe

mal

eat

bord

erA

lph

am

ale

atbo

rder

Opp

ortu

nis

tic

adva

nce

&re

trea

t

Mal

e-m

ale

con

flic

t

Fem

ale-

fem

ale

con

flic

t

Her

din

gIn

terg

rou

pse

xO

utc

ome

On

esi

deB

oth

side

sTo

tal

On

esi

deB

oth

side

sTo

tal

Aff

init

ive

inte

ract

ion

On

esi

deB

oth

side

sTo

tal

Init

iati

onM

atin

gW

inn

eran

dlo

ser

Un

cert

ain

Bir

thS

.(n

515

)4

1014

311

149

119

17

512

00

150

Mat

ing

S.(

n5

14)

211

134

1014

913

104

66

126

43

11A

vera

gera

teof

two

S.

2172

9324

7296

6283

6618

4538

83—

Dif

fere

nce

betw

een

S2

NS

—N

SN

SN

SN

SN

S15

.8**

*(z

)(z

)(z

)(z

)(z

)(z

)(X

2)

1S

eefo

otn

ote

1,Ta

ble

2a.

2z

(pro

port

ion

test

):N

S,n

otsi

gnifi

can

t;X

2(c

rost

ab):

***

P,

0.00

1.

466 Q.-K. ZHAO

Competition for food was mediated by therelative dominance rank between groupsinvolved, just like the interaction amongindividuals within a group (Deng and Zhao,1987). The intergroup dominance, however,was not simply determined by demographicvariables such as adult females, adult males,total adults, or group size. The rank-rever-sal between groups A and B showed thatmore subtle factors such as past relationsamong males and experiences of femalesbeing with the males might influence theintergroup rank order, as in rhesus ma-caques (Gabow, 1972) and red colobus (Struh-saker and Leland, 1979; but see Cheney,1987).

Forewarning is very likely to function asan advertisement of the group’s presencethat may discourage an intruder in the birthseason, as similar behavior does in mantledhowler monkeys (Chivers, 1969) and capu-chins (Terborgh, 1983). Because most ma-caques seem to lack specialized intergroupcalls (Cheney, 1987), the forewarning behav-ior observed in Tibetan macaques may bestimulated by the high intensity of competi-tion for human food in the tourist season.

With the change in sexual activity (ofintra- and intergroup) from absence in thebirth season to presence in the mating sea-

son, the proportion of close encounters intotal interactions increased, intergroupdominance became obscure, the asymmetryof forewarning behaviors (resident groupsover intruders) disappeared, the rate ofbranch shaking decreased, and sometimesintergroup cohesion (as described in foot-note 1) appeared. The results suggest thatthe force of exclusion in the birth season wasoutweighed by sexual attractiveness be-tween individuals from different groups, andreversely in the next birth season, that is,the two contended forces were in a dynamicequilibrium across the seasons. This observa-tion holds those on the motive of sexualattractiveness between unfamiliar animalsfrom different groups (Japanese macaques:Enomoto, 1974; Sugiyama, 1976; rhesus ma-caques: Brereton, 1981; Hausfater, 1992;Tibetan macaques: Zhao, 1993), and unfamil-iar opposite sexes of humans (Shepher, 1971;Wolf and Huang, 1980) though the attractive-ness was not clearly documented as a con-tended force of the competition in previousstudies.

In conspicuous contrast with the conflictsbetween young immigrant males and theresident alpha male in the mating season,which produce a lot of wounds in the youngadult immigrants (Zhao, 1994a), and withfemale-female conflicts between groups, theintergroup male-male conflict was only aritualized show of power, which never pro-duced any wounds. Forewarning display inthis species is probably the same kind ofshow. Both the male-male conflict and theforewarning behavior are like by-products offemale choice in mating activities within afemale-banded group (Wrangham, 1980), do-ing so to gain the favor of females within thegroup.

Interrelatedly, males herding females inclose encounters is typically observed inbaboons (Stoltz and Saayman, 1970; Hamil-

TABLE 3. Intergroup dominance and group composition observed mainly in the 1991 birth season

Winner-loser1 A-D1 B-D1 D-D1 E-D1 A-B D-B D-A(n) (1) (6) (2) (2) (1) (3) (1)Number of adult females 7–6 8–6 25–6 10–6 7–8 25–8 25–7Number of adult & subadult males 4–4 6–4 9–4 5–4 4–6 9–6 9–4Number of adults 11–10 14–10 34–10 15–10 11–14 34–14 34–11

(Intergroup rank order: D . A . B . D1)1 No reverse was observed in the sampling period; the interaction between D and A occurred on December 1, 1989.

TABLE 4. Changes in trail-range use in groups B andD, and the trail sections ranged by fissioned minor

groups D1 and D2

Group Beginning1

1991/1992

Extendedin

Givingup

Group B b c d f g h a e b hGroup D c d g h i b g2 i 2

Group D1 c d g2 h e fGroup D2 c i 2

1 Data collected in 1986 for groups B and D; groups D1 and D2were fissioned from group D in October 1988 and late August1991 respectively.2 Trail-sections given up by group D before the fissions.

467INTERGROUP INTERACTIONS IN TIBETAN MACAQUES

ton et al., 1975; Cheney and Seyfarth, 1977;Cowlishaw, 1995). Such herding is thoughtto bar sexually receptive females from ac-cess to males in other groups (Cheney, 1987).On the other hand, sexual weight dimor-phism has also been hypothesized as animportant factor influencing male herdingbecause herding is less common in lesssexually dimorphic species such as Japa-nese macaques and vervets (Cheney, 1981,1987). The nearly identically high rates ofherding observed during intergroup encoun-ters in the birth and mating seasons (80%and 85% respectively) rules out the cause forherding hypothesized for baboons, i.e. pre-venting sexually receptive females accessmales in other group (Cheney, 1987) becauseno receptive females existed in the birthseason. On the other hand, when the weightdimorphism reduced from to 1.46 (M/F) inlater winter to 1.16 in late autumn (themating season; Zhao, 1994b), the rate ofherding showed no decline. This arguesagainst the link suggested between herdingand size dimorphism (Cheney, 1981, 1987).Possibly, herding in Tibetan macaques isselected for a more subtle factor such assmoothing the males next intergroup trans-fer by doing so to gain the favor of femalesfrom the opposite group during encounters.Thus, attempting to gain the favor of fe-males both within and outside the group,high-ranking males’ reproductive strategiesare embodied in all of the performances inintergroup interactions.

Consistent with reports on other species,the young natal males (rhesus: Hausfater,1972; Japanese macaques: Sugiyama, 1976),and the subordinate males (vervets: Cheneyand Seyfarth, 1983) were the most activeparticipants in intergroup encounters, espe-cially in affinitive interactions. It is alsopossible to consider that affinitive perfor-mances of peripheral males at the border areessential for their coming or next transfer.Thus, different tactics are used by the twoclasses of males in preparing their coming ornext transfers in Tibetan macaques. Maleintergroup transfer is closely related to rais-ing the mating opportunity for the youngimmigrants, and to avoiding intermale com-petition for the secondary transfers that aremiddle-aged and old (Zhao, 1993, 1994a). In

addition to the role of couriers of genes (e.g.,Nozawa et al., 1975) and culture (reviewedby Nishida, 1987; Zhao and Deng, 1992),transferring males may also go between thegroups in reallocation of ranges, as thatdescribed in footnote 1.

Females were the most serious fighters insevere conflicts between groups, especiallybetween two groups (D-D1 and D-D2) fis-sioned in the recent past. This may well berelated to the fact that after fission the twoparts shared the same home range, and theminority was always forced to forage in thepoorer parts. Interestingly, similar trend isalso reported for Barbary macaque groupsfissioned in the recent past (Prud’Homme,1991), but not for Japanese monkeys(Koyama, 1970) and rhesus (Missakian,1973), in which the relations between re-cently fissioned groups are initially unag-gressive. In addition to the extra-develop-ment of alloparenting or male-infant care-taking [Tibetan macaques (TM): Deng andZhao, 1996; Barbary macaques (BM): Whit-ten, 1987.] and male-infant-male interac-tions (TM: Zhao, 1996a; BM: Deag, 1980;Taub, 1980), this extrahostility between fis-sioned groups is likely another example ofecological factors shaping social behavior(Zhao, 1996b). Of the factors, what we knowis that Tibetan (Zhao et al., 1991) and Bar-bary (Menard and Vallet, 1988) macaquesunder poor foraging conditions feed more onfoliage than other macaques, as indicated bytheir body weights (Clutton-Brock and Har-vey, 1977; Sailer et al., 1985)—the Tibetan(Zhao, 1994b) and Barbary (Harvey et al.,1987; Fa, 1986) macaques are the largestspecies of their genus.

Under the given ecological conditions, fis-sion, which occurred only in the top-rankingand largest group twice during the studyperiod, is likely an outcome of balance be-tween the cost of competition for food withinthe group, and the benefit of cooperativeaccess to food resources for animals in themother group. That is, when the former isoutweighed by the latter, fission occurs toreduce the intra-group competition, but doesnot reduce the ability of the mother group toaccess the best feeding sites.

However, mate competition is very likelyto be proximately involved with group fis-

468 Q.-K. ZHAO

sion because both observed fissions occurredin the mating season. Tibetan macaques cantemporarily form a ‘‘far-peripheral adult sub-group’’ (FAS) in the mating season as a‘‘space-segregation’’ tactic for mating by theloser among males and females (Zhao, 1993),thus fission may also be viewed as thedevelopment of this tactic for the daughtergroup.

In summary, this study first fully demon-strates the theory of sexual selection (Dar-win, 1871; Trivers, 1972) with quantitativedata on behavioral differences between fe-males and males in intergroup interactions—females and males do maximize reproduc-tive values in different way. Second, sexualattractiveness among groups is integrated/introduced into the model of the female-banded society of primates as a force con-tended against food competition. Third, thestudy groups demographic data collectedwith two group-fissions afford an ecologicalexplanation of group fission.

The decline of sizes in groups A, B and Dmay be related to poaching that became aproblem after 1987, though it stopped fromtime to time in the study period. (Zhao,1994a).

ACKNOWLEDGMENTS

In the field, this study was conducted withpermission and assistance from the Adminis-trative Commission of Mt. Emei and Xian-feng Temple. I am grateful to reviewers forvaluable comments on the manuscript andto Dr. M. Cartmill and Dr. E.J.E. Szathmaryfor their critical editorial work. Specialthanks are given to Drs. B. Foster and L.T.Nash for their help in the department ofAnthropology, Arizona State University,when I was preparing the long-term project.

LITERATURE CITED

Bertrand M (1969) The Behavioral Repertoire of theStumptail Macaques. Basel: S. Karger.

Brereton A (1981) Inter-group consorting by a free-ranging female rhesus monkey (Macaca mulatta).Primates 22:417–423.

Cheney DL (1981) Inter-group encounters among free-ranging vervet monkeys. Folia Primatol. 35:124–146.

Cheney DL (1987) Interactions and relationships be-tween groups. In BB Smuts, DL Cheney, RM Seyfarth,RW Wrangham, and T Struhsaker (eds.): PrimateSociety. Chicago: University of Chicago Press, pp.267–281.

Cheney DL, and Seyfarth RM (1977) Behavior of adultand immature male baboons during inter-group en-counters. Nature 269:404–406.

Cheney DL, and Seyfarth RM (1983) Non-random dis-persal in free-ranging vervet monkeys: Social andgenetic consequences. Am. Nat. 122:392–412.

Chivers DJ (1969) On the daily behavior and spacing ofhowling monkey groups. Folia Primatol. 10:48–102.

Clutton-Brock TH, and Harvey PH (1977) Species differ-ences in feeding and ranging behaviour in primates.In TH Clutton-Brock (ed.): Primate Ecology. London:Academic Press, pp. 557–584.

Cowlishaw G (1995) Behavioral patterns in baboongroup encounters: The role of resource competitionand male reproductive strategies. Behaviour 132:75–86.

Darwin C (1871) The Descent of Man, and Selection inRelation to Sex. London: John Murray.

Deag, JM (1980) Interactions between males and un-weaned Barbary macaques: Testing the agonistic buff-ering hypothesis. Behaviour 75:54–81.

Deng Z-Y, and Zhao Q-K (1987) Social structure in a wildgroup of Macaca thibetana at Mt. Emei, China. FoliaPrimatol. 49:1–10.

Deng Z-Y, and Zhao Q-K (in press) Alloparenting fornewborns of Tibetan macaques.ActaAnthropol. Sinica.15:159–165. (in Chinese with English abstract).

Enomoto T (1974) The sexual behavior of Japanesemonkeys. J. Hum. Evol. 3:351–372.

Fa JE (1986) Use of Time and Resources by ProvisionedTroops of Monkeys. Basel: S. Karger.

Gabow SL (1972) Dominance order reversal betweentwo groups of free-ranging rhesus monkeys. Primates14:215–223.

Harvey HP, Martin RD, and Clutton-Brock TH (1987)Life history in comparative perspective. In BB Smuts,DL Cheney, RM Seyfarth, RW Wrangham, and TStruhsaker (eds.): Primate Socieities. Chicago: Chi-cago University Press, pp. 181–196.

Hamilton WJ, Buskirk RE, and Buskirk WH (1975)Chacma baboon tactics during intertroop encounters.J. Mammal. 56:857–870.

Hausfater G (1972) Intergroup behavior of free-rangingrhesus monkeys (Macaca mulatta). Folia Primatol.18:78–107.

Kawanaka K (1973) Intergroup relations among Japa-nese monkeys. Primates 14:113–159.

Koyama N (1970) Changes in dominance rank anddivision of a wild Japanese monkey troop in Arashi-yama. Primates 11:335–390.

Lindburg DG (1977) Feeding behaviour and diet ofrhesus monkeys in a Siwalik forest in North India. InTH Clutton-Brock (ed.): Primate Ecology. London:Academic Press, pp. 223–249.

Menard N, and Vallet D (1988) Disponibilites et utilisa-tion des ressources par le magot (Macaca sylvanus)dans differents milieux en Algerie. Rev. Ecol. (TerreVie) 43:201–250. (English abstract)

Missakian EA (1973) The timing of fission among free-ranging rhesus monkeys. Am. J. Phys. Anthropol.38:621–624.

Mitani JC, and Rodman PS (1979) Territoriality: Therelation of ranging patterns and home range size todefendability, with an analysis of territoriality amongprimate species. Behav. Ecol. Sociobiol. 5:241–251.

Nishida T (1987) Local tradition and cultural transmis-sion. In BB Smuts, DL Cheney, RM Seyfarth, RWWrangham, and T Struhsaker (eds.): Primate Societ-ies. Chicago: University of Chicago Press, pp. 462–474.

Nozawa K, Shotake T, Ohkura Y, Kitajima M, andTanabe Y (1975) Genetic variations within and be-tween troops of Macaca fuscata. In S Kondo, A Ehara,

469INTERGROUP INTERACTIONS IN TIBETAN MACAQUES

M Kawai (eds.): Contemporary Primatology. Basel: S.Karger, pp. 75–89.

Prud’Homme J (1991) Group fission in a semifree-ranging population of Barbary macaques (Macacasylvanus). Primates 32:9–22.

Sailer LD, Gaulin SJC, and Kurland JA (1985) Measur-ing the relationship between dietary quality and bodysize in primates. Primates 26:14–27.

Shepher J (1971) Mate selection among second genera-tion kibbutz adolescents and adults: Incest avoidanceand negative imprinting. Arch. Sex. Behav. 1:293–307.

Southwick CH, Beg MA, and Siddigi MR (1965) Rhesusmonkeys in North India. In J Dovore (ed.): PrimateBehavior. New York: Holt, Rinehart, and Wintson, pp.111–159.

Stoltz LP, and Saayman GS (1970) Ecology and behaviorof baboons in the northern Transvaal. Ann. TransvaalMus. 26:99–143.

Sugiyama Y (1976) Life history of male Japanese mon-keys. In JS Rosenblatt, RA Hinde, E Shaw, and C Beer(eds.): Advances in the Study of Behavior. New York:Academic Press, pp. 155–284.

Taub DM (1980) Testing the ‘agonistic buffering’ hypoth-esis. Behav. Ecol. Sociobiol. 6:187–197.

Terborgh J (1983) Five New World Primates: A Study onComparative Ecology. Princeton: Princeton Univer-sity Press.

Trivers RL (1972) Parental investment and sexual selec-tion. In B Campbell (ed.): Sexual Selection and theDescent of Man, 1871–1971. Chicago: Aldine, pp.136–179.

Waser PM (1976) Cecocebus albigena: Site attachment,avoidance, and intergroup spacing. Am. Naturalist110:911–935.

Whitten PL (1987) Infants and adult males. In BBSmuts, DI Cheney, RM Seyfarth, RW Wrangham, andT Struhsaker (eds): Primate Societies. Chicago: Chi-cago University Press, pp. 343–357.

Wolf AP, and Haung C-S (1980) Marriage and Adoptionin China. 1845–1945. Stanford: Stanford UniversityPress.

Wrangham RW (1980) An ecological model of female-bonded primate groups. Behaviour 75:269–299.

Zhao Q-K (1993) Sexual behavior in Tibetan macaques(Macaca thibetana) at Mt. Emei, China. Primates34:431–444.

Zhao Q-K (1994a) Male mating competition and inter-group transfer in Tibetan macaques (Macaca thi-betana) at Mt. Emei, China. Primates 35:57–68.

Zhao Q-K (1994b) Seasonal change in body weight ofMacaca thibetana at Mt. Emei, China. Am. J. Prima-tol. 32:223–226.

Zhao Q-K (1994c) A study on semi-commensalism ofTibetan macaques at Mt. Emei, China. Rev. Ecol.(Terre Vie) 49:256–271.

Zhao Q-K (1994d) Birth timing shift with altitude andits ecological implications for Macaca thibetana at Mt.Emei. Oecologia Montana 3:24–26.

Zhao Q-K (1996a) Male-infant-male interactions in Ti-betan macaques. Primates 37:135–143.

Zhao Q-K (1996b) Etho-ecology of Tibetan macaques atMount Emei, China. In JE Fa and DG Lindburg (eds):Evolution and Ecology of Macaque Societies. Cam-bridge: Cambridge University Press, pp. 263–289.

Zhao Q-K, and Deng Z-Y (1988) Macaca thibetana at Mt.Emei, China: III: Group composition. Am. J. Primatol.16:269–273.

Zhao Q-K, and Deng Z-Y (1992) Dramatic consequencesof food handouts to Macaca thibetana at Mt. Emei,China. Folia Primatol. 58:24–31.

Zhao Q-K, Deng Z-Y, and Xu J-M (1991) Natural foodsand their ecological implications for Macaca thibetanaat Mt. Emei, China. Folia Primatol. 57:1–15.

Zhao Q-K, Xu J-M, and Deng Z-Y (1989) Climate,vegetation and topography of the slope habitat ofMacaca thibetana at Mt. Emei, China. Zool. Res.10:92–100.

470 Q.-K. ZHAO