The formation of red colobus–diana monkey associations under predation pressure from chimpanzees R O N A L D N O E>    R E D O U A N B S H A R Y Max-Planck-Institut fuX r Verhaltensph’siologie, SeeWiesen, Postfach 1564, 82319 Starnberg, German’

SUMMARY

It is generally assumed that most primates live in monospecific or polyspecific groups because group living provides protection against predation, but hard evidence is scarce. We tested the antipredation hypothesis with observational and experimental data on mixed-species groups of red colobus (Procolobus badius) and diana monkeys (Cercopithecus diana) in the Taı$ National Park, Ivory Coast. Red colobus, but not diana monkeys, are frequently killed by cooperatively hunting chimpanzees. Association rates peaked during the chimpanzees’ hunting season, as a result of changes in the behaviour of the red colobus. In addition, playbacks of recordings of chimpanzee sounds induced the formation of new associations and extended the duration of existing associations. No such effects were observed in reaction to control experiments and playbacks of leopard recordings.

The antipredation hypothesis gained wide acceptance as a result of observational and comparative studies of diurnal primates living in monospecific groups (van Schaik 1983 ; van Schaik et al. 1983 ; van Schaik & van Noordwijk 1985 ; van Hooff & van Schaik 1992 ; Janson 1992). It has also been tested for a number of mixed-species groups (Struhsaker 1981 ; Gautier-Hion et al. 1983 ; Terborgh 1983 ; Cords 1987, 1990 ; Struhsaker & Leakey 1990 ; Chapman & Chapman 1996). However, conclusive evidence supporting the antipredation hypothesis has neither been reported for monospecific, nor for polyspecific groups (Cheney & Wrangham 1987 ; Isbell 1994). The formation of an association with individuals of another species can be an alternative to a larger monospecific group when the species in the mixed group are subject to the same predators. A mixedspecies group has several potential advantages. First, food competition can be considerably lower when the diets show little overlap. Secondly, partner species can have complementary skills in antipredation defence, or use forest strata in such a way that they ‘ shield ’ each other (Gautier-Hion et al. 1983). Third, predators may show preference for one of the species (cf. FitzGibbon 1990 for two gazelle species). In this paper we test the antipredation hypothesis for mixed groups formed by two monkey species : the red colobus (Procolobus badius) and the diana monkey (Cercopithecus diana). In the Taı$ National Park, Ivory Coast, each red colobus group forms regular associations with a specific diana partner-group with which it shares a communal range (Ho$ ner et al. 1997). The association rates are considerably higher than would be expected on the basis of chance encounters alone, but the groups are not permanently together (Holenweg et al. 1997). Our data do not reveal any foraging-related advantages for

1. I N T R O D U C T I O N

There are three principal explanations for the occurrence of polyspecific associations among forestliving primates : chance encounters of groups of mutually tolerant species ; improved foraging efficiency ; and reduced predation. Primate groups can meet by chance when they are attracted to resources in the same location or cross each other’s path while travelling through their range (Waser 1982, 1984 ; Whitesides 1989 ; Cords 1990 ; Oates & Whitesides 1990). Species with different diets have little reason to avoid each other. The remaining two explanations imply that members of at least one species are directly attracted to groups of the other species. The improvement of foraging efficiency is one possible cause of attraction (Cords 1987, 1990 ; Garber 1988 ; Peres 1992 a). One species may, for example, guide the other to momentarily available food sources, or may flush mobile prey from the vegetation. According to the antipredation hypothesis the risk of predation would be lower in the presence of an allospecific group than in it’s absence. Basically the same hypothesis has been put forward to explain why many primate species live in larger groups than the smallest possible reproductive unit (Alexander 1974 ; van Schaik 1983 ; van Schaik & van Hooff 1983 ; Terborgh & Janson 1986 ; Dunbar 1988). Groups are assumed to provide ‘ safety in numbers ’, which is the combined effect of dilution, predator confusion, early warning and communal defence (Williams 1966 ; Hamilton 1971 ; Pulliam & Caraco 1984). The risk of being detected increases with group size, but this is more than compensated for by the dilution effect for predators that kill only a fraction of a group per encounter (Turner & Pitcher 1986 ; Inman & Krebs 1987). Proc. R. Soc. Lond. B (1997) 264, 253–259 Printed in Great Britain

253

" 1997 The Royal Society

254

R. Noe$ and R. Bshary

Red colobus–diana associations

either species (Wachter et al. 1997). On the contrary, the main constraint on associations is probably the incompatibility of foraging patterns, because the two species have a very low diet overlap. This leaves antipredation defence as the most likely function of the association behaviour. The variation in association rates between red colobus and diana monkeys gave us a unique opportunity to prove the existence of a direct causal connection between predation pressure and association behaviour. We hypothesized that sufficiently strong natural or experimentally simulated fluctuations in predation pressure would cause noticeable changes in association behaviour, assuming that the monkeys can monitor the changes in the level of risk. There are four major predators of monkeys in the Taı$ Park : crowned hawk eagles (Stephanoaetus coronatus), leopards (Panthera pardus), chimpanzees (Pan troglod’tes Šerus) and humans. We concentrated on predation by chimpanzees for the following reasons. First, chimpanzees frequently kill red colobus, but only rarely hunt diana monkeys (Boesch & Boesch 1989). We could thus limit our hypotheses and tests to one species only. Second, chimpanzees show strong seasonality in hunting activity (Boesch 1994 a, b) in contrast to the other predators. Third, because the range of a chimpanzee community is about 50 times the size of the range of a colobus group, encounters between chimpanzees and a particular red colobus group are relatively rare. In Taı$ , most chimpanzee hunts involve groups of males (Boesch & Boesch 1989 ; Boesch 1994 a, b). Hunting males are silent, but otherwise they regularly engage in loud displays. These displays and other loud vocalizations inform the monkeys about the presence and approximate composition of chimpanzee parties in their vicinity. We expected therefore that chimpanzee predation had the necessary attributes to make a flexible, momentary adaptation to changes in risk by the prey possible. The above mentioned characteristics of the chimpanzee–red colobus interaction are specific for the Taı$ forest (cf. Stanford et al. 1994 ; Stanford 1995 for Gombe, Tanzania). 2. M E T H O D S (a) Study groups Our main subjects were two combinations of one Colobus badius and one Cercopithecus diana group (Bad 1¬Dia 1 and Bad 2¬Dia 2). For playback experiments we used these groups and seven unhabituated C. badius groups with their C. diana partner groups. Unhabituated red colobus groups were identified on the basis of their home range. In the Taı$ forest, ranges of neighbouring red colobus groups show little overlap (Bshary 1995 ; Ho$ ner et al. 1997). The study was conducted between March 1991 and December 1994. (b) Association criteria For the analysis of seasonal fluctuations we calculated association rates by dividing the time the groups were seen ‘ in association ’ by the total observation time. Two groups belonging to different species were considered associated when their members were at least partially intermingled, i.e. the imaginary polygons connecting the outermost members Proc. R. Soc. Lond. B (1997)

of each group overlapped. During the first nine months of 1991 we used a less strict criterion : two groups were considered to be associated when the distance between the imaginary polygons connecting the outermost members of each group was less than 25 m. We estimated the time the study groups had been partially intermingled during those nine months by using a correction factor calculated on the basis of data gathered during eight months ( July 1992–March 1993) in which both criteria were used simultaneously. The error thus introduced is small, because the association rates calculated on the basis of the two criteria differ only slightly. We could not always observe the unhabituated groups used in the experiments well enough to establish whether their members were intermingled or not. In some cases, which are indicated in the text, we therefore considered two groups associated when individuals belonging to different species were observed at less than 50 m distance from each other. (c) Seasonality We used periods of three months for the analyses of seasonality, because the hunting season of the chimpanzees (September–November) and the dry season (December– February) each last approximately three months. The hunting season is a period of heavy rainfall. The remaining period was divided in two periods of three months each : rain 1 (March–May) and rain 2 ( June–August). During the latter period the average party size of chimpanzees is lower than in the months before or after (Doran 1989). We tested for seasonality in association rates (total time in association}total observation time) by calculating autocorrelations of logtransformed rates per season corrected for linear trend, using the  statistical package. A significant correlation with a lag of four seasons means that association rates fluctuate according to a yearly cycle. (d) Approach-leave analysis A group was considered to approach or leave an allospecific group when the core of the group moved over more than a distance of 50 m and the association state changed between ‘ not intermingled ’ and ‘ intermingled ’. (e) Playback experiment 1 : prolongation of associations Period : July–August 1993. Each of the nine combinations of a red colobus and a diana group was subjected to each type of playback stimulus only once within this period. The groups were intermingled at the start of the experiment. Groups were considered to remain in association as long as the distance between the nearest members of different groups observed did not exceed 50 m. Time of day of the playback : between 06:45 and 07:30. Position of the loudspeaker : approximately equidistant from the cores of both groups at about 50 m. Sound recordings : three different recordings of a diesel generator, three different recordings of leopard calls from East Africa (source : band 13, British Library of Wildlife Sounds ; BLOWS), three different recordings from East African chimpanzees (source : disk N° MP 25224 (recorded by V. Reynolds, Budongo Forest, Uganda), BLOWS). A field assistant imitated the low pitched chimpanzee drumming simultaneously to the playback. No recording was used more than three times. We were confident that the calls recorded in East Africa provided the stimulus we intended, as diana monkeys in our area reacted with the alarm calls and long calls typically given in reaction to leopards to recordings from East African leopards (Zuberbu$ hler et al. 1997). The

Red colobus–diana associations reactions to calls from East African and West African chimpanzees (used in experiment 2) were indistinguishable. Broadcasting equipment : Panasonic RX-CS780 with an 80 W amplifier and a 50–15 000 Hz range. Broadcasting pattern : 20 s sound, 10 s break, 10 s sound. ( f ) Playback experiment 2 : formation of associations Period : May–July 1994. Groups, broadcasting equipment and pattern, as for experiment 1. Position of groups : " 100 m from the border of the communal range, cores of groups " 100 m from each other. Position of the loudspeaker : " 100 m and approximately equidistant from the cores of both groups. Sound recordings : generator and leopard as in experiment 1. Chimpanzees : four different recordings from Taı$ (source British Broadcasting Corporation, Natural History Unit, recorded for the television series Trials of Life), chimpanzee drumming as in experiment 1. 3. R E S U L T S (a) Predation pressure by chimpanzees

The red colobus are estimated to have a density of between 2.35 groups}km# (transect method ; Holenweg et al. 1996) and 2.47 groups}km# (based on the home range size and range overlap of two neighbouring study groups ; Bshary 1995). Our study groups had 70 and 75 members. This gives an estimated number of 63–67 red colobus groups, or 4442–5002 individuals, within the 27 km# range (Boesch & Boesch 1989) of a single chimpanzee community. The antipredation strategy of the red colobus will depend on the number of encounters with parties of male chimpanzees and the number of actual attacks. Assuming there is a single party with enough males to stage a cooperative hunt per chimpanzee community at any time, a colobus group would on average have 2.2–2.3 high-risk encounters}month (based on encounter rates given in

Ronald Noe$ and Redouan Bshary

Boesch (1994 a) and on our density estimates). The number of actual attacks is estimated to be 0.71–0.74}colobus group}month in the hunting season (September–November) and 0.15–0.16}colobus group}month during the rest of the year (calculated on the basis of attack rates in Boesch (1994 a) and on our estimates of colobus group densities). On average 0.57 colobus are killed per attack (Boesch & Boesch 1989). (b) Seasonal variation in hunting pressure and association behaviour

We found (figure 1) a marked seasonal variation in association rates. A significant autocorrelation for a time lag of four quarters confirms that the association rate fluctuates according to a yearly cycle (based on 15 seasons for combination Bad 1¬Dia 1, autocorrelation coefficient 0.541, p ! 0.05, two-tailed test for correlation coefficients ; Sokal & Rohlf 1995). The observation period for the combination Bad 2¬Dia 2 was too short to make this analysis, but the pattern is similar (figure 1). For both groups the association rate was always higher during the hunting season than during the preceding season, and, with one exception, higher than during the following season. Association rates were particularly low during the season ‘ rain 2 ’, the season in which the average party size of the chimpanzees is low. The relative contribution of red colobus to both the formation and the maintenance of associations was higher during hunting seasons than during the rest of the year (table 1). (c) Experiment 1 : playbacks to existing associations

Experiment 1 demonstrated that a short bout of chimpanzee sounds in the early morning sufficed to extend the duration of existing associations consider-

80

Association rate %

70

60

50

40

30 rain 1 rain 2 hunt 1991

dry

rain 1 rain 2 hunt 1992

dry

rain 1 rain 2 hunt 1993

dry

rain 1 rain 2 hunt 1994

Figure 1. Association rates of two red colobus–diana combinations. Association rates are expressed as percentages of total observation time during periods of 3 months for two pairs of red colobus–diana groups. Based on 10881 h of observation (at least one observer with either group) for the combination Bad 1¬Dia 1 (solid line) collected between March 1991 and November 1994 and on 4626 h for Bad 2¬Dia 2 (dotted line) collected between September 1992 and November 1994. Solid circles denote ‘ hunting season ’ of chimpanzees. Proc. R. Soc. Lond. B (1997)

255

R. Noe$ and R. Bshary

256

Red colobus–diana associations

Table 1. Responsibility for maintenance of associations (Data for groups Bad I and Dia I for 1992–94. Hunting season: September – November. Expected values in brackets. Two-tailed Gtests.) approach

hunting seasons other seasons

(d) Experiment 2 : playbacks to separated partner groups

leave

red colobus

diana

96

55

(82.4)

(68.6)

183

177

(196.6)

(163.4)

151

360

279 232 G = 7.04, p = 0.0080

red colobus

diana

15

157

(27.1)

(144.9)

77

335

(64.9)

(347.1)

chimpanzee sounds : empty tape-chimpanzee : p ! 0.01 ; generator-chimpanzee : p ! 0.05. The differences between other stimulus pairs were not significant.

172

412

92 492 G = 9.99, p = 0.0016

ably (table 2). An overall test showed that at least some stimuli induced significantly different reactions (Friedman’s two-way ANOVA by ranks ; Conover 1980 ; T ¯ 3.29, k ¯ 3, k ¯ 24 ; two-tailed test p ! # " # 0.05). Subsequent pairwise comparisons revealed significant differences between the control stimuli and the

A short stimulus of chimpanzee sounds sufficed to induce the red colobus to move towards their diana partners (figure 2). As in experiment 1, a leopard stimulus did not have a significant effect. We tested the outcomes under the assumption that the two groups behaved independently. The overall statistical test was significant for red colobus (number of approaches compared with staying or leaving, Cochran’s Q ¯ 17.22, d.f. ¯ 3, p ! 0.001)), but not for diana (Q ¯ 6.17, d.f. ¯ 3, p " 0.10). Cochran’s test for multiple comparisons (Bortz et al. 1990) revealed that for red colobus the effect of the chimpanzee stimulus differed significantly from the effect of the other three stimuli (empty tape-chimpanzee p ! 0.05 ; generator-chimpanzee p ! 0.02, leopard-chimpanzee p ! 0.01). The differences between other stimulus pairs were not significant.

Table 2. Pla’back experiment 1 : duration of association after earl’-morning stimulus (Scores reflect numbers of 15 min periods the two groups stayed together after the playback. & 38 : till end of observation at 17:00. Combinations 1 and 2 are the habituated study groups.) red colobus–diana combination empty tape

generator

leopard

chimpanzee

1 2 3 4 5 6 7 8 9

& 38 11 25 12 14 & 38 24 24 & 38

& 38 & 38 15 20 & 38 & 38 5 & 38 35

& 38 & 38 35 33 & 38 & 38 & 38 28 & 38

16 & 38 5 1 & 38 & 38 & 38 12 5

Figure 2. Playback experiment 2 : formation of associations. Position of the two partner groups 30 min after the playback experiment. Distance between cores of groups at start of experiment : " 100 m. Hatched symbol : group did not move ; open symbol : group moved " 50 m away from the partner group ; solid symbol : group approached the partner group and intermingled. Proc. R. Soc. Lond. B (1997)

Red colobus–diana associations After a chimpanzee experiment the gap between the groups was closed by the red colobus rather than by the diana group (Fisher Exact p ¯ 0.07, two-tailed). One cannot say, however, whether diana monkeys would have approached in case the red colobus would not have moved. 4. D I S C U S S I O N

Red colobus associate more with diana monkeys than with any of the other five monkey species in our study area (Ho$ ner et al. 1997). This preference is probably based on a combination of factors, including low food competition, compatibility of home range size and the total vigilance level of the diana group. The latter is in turn due to a high individual vigilance level, a relatively large group spread and specific foraging habits (Bshary 1995 ; McGraw 1996 ; Bshary & Noe$ , submitted). Our data suggest that red colobus seek the company of diana monkeys to reduce the predation risk due to chimpanzees. The colobus are likely to profit from diana alarm calls for other predators as well. The absolute contribution of diana monkeys to the formation of associations was almost as high as that of the red colobus (table 1). This indicates that they were neither parasitized by the colobus, nor profited passively. A diana group might gain because red colobus shield them against attacks by crowned eagles, which prey upon both Cercopithecus and Colobus monkeys (Struhsaker & Leakey 1990). By forming an association with red colobus the group size jumps from 25 to 100. This ‘ dilution effect ’ is important during the attacks of surprise hunters, notably eagles, and could explain the high association rates during the dry season when the foliage is thin. Although diana monkeys rarely fall victim to chimpanzees, they are not immune and give alarm calls for chimpanzees (Boesch & Boesch 1989 ; Zuberbu$ hler et al. 1997). Diana monkeys are associated with either red or black-and-white colobus most of the time (Bshary 1995). These large colobus form the bulk of the prey of the Taı$ chimpanzees (Boesch & Boesch 1989). Diana monkeys might therefore be protected by the presence of preferred prey (cf. FitzGibbon 1990). One can pose two fundamentally different questions. (1) Which primary ultimate factor initially caused the selection of association behaviour ? (2) Under which conditions do groups associate ? With ‘ association behaviour ’ we refer to behaviour that results in the formation and maintenance of associations with allospecifics that would not occur as a result of chance encounters. The first question refers to the origin of this behaviour in each of the two species or their ancestors. It is very unlikely that more than a single novel environmental factor triggered the initial selection. A likely factor for both species is predation pressure by crowned hawk eagles or similar birds of prey (Struhsaker 1981 ; Struhsaker & Leakey 1990 ; cf. van Schaik & Ho$ rstermann 1994). Present-day red colobus and diana monkeys use conditional association strategies. Only species that form permanent associations (e.g. olive colobus : Oates & Whitesides 1990 ; Proc. R. Soc. Lond. B (1997)

Ronald Noe$ and Redouan Bshary

257

tamarins : Peres 1992 a, b) may use an unconditional strategy. Ancestral association strategies must also have been conditional, because a sudden jump to permanent associations is unlikely. We assume that associations in reaction to surprise hunters, like eagles, are conditional upon foraging constraints only. A defence strategy cannot be conditional upon predator pressure in the absence of clues that indicate fluctuations in predation risk. Factors that played a role later in evolutionary time may have further shaped the strategy of each species. For example, a predator species that threatened each monkey group at irregular intervals could have caused selection for a strategy conditional upon the momentary predation pressure, assuming such fluctuations could be monitored in some way. Examples of predators with such characteristics are human hunters practising shifting ‘ slash and burn ’ agriculture and, as we discuss below, chimpanzees. The second question can be answered by the study of present-day causes. We assume that the tendency to associate depends on the total cost-benefit budget. Many different factors may contribute to the fitness effect of a change in association state. In the case of red cyolobus–diana monkey associations important budget items are costs due to the incompatibility of foraging patterns and benefits due to reduced mortality. The pattern of seasonal fluctuations of the cost-benefit budget is bound to be complicated because four predator species and several hundred food species are involved. We expected to see a measurable correlation only between association rates and environmental factors that (1) have a strong impact, (2) show strong seasonal fluctuations and (3) can be monitored by the prey species. The hunting pressure of chimpanzees on colobus proved to be such a factor. The availability of specific food items shows strong, easily noticeable fluctuations, but the abundance of a single item is not likely to have a strong impact, because it is swamped in the broad diets of both species. However, we cannot exclude the possibility that general seasonal fluctuations in food availability had an effect. The analysis of the seasonal fluctuation in association rates shows that red colobus and diana groups correlates with the hunting pressure of the chimpanzees. During the season ‘ rain 2 ’, food constraints cause the chimpanzees to forage in small parties (Doran 1989). It is conceivable that monkeys do not form mixed groups during the same season for a similar reason. Competition between red colobus and diana is negligible, but scarce resources could aggravate the problem of incompatibility of foraging patterns. When resource levels would explain the fluctuations in hunting activity of the chimpanzees and the association rates of the monkeys independently, the question of which crucial factor causes the red colobus to seek associations with diana monkeys would remain unanswered. Fluctuations in the resources of diana monkeys (ripe fruit and insects ; Wachter et al. 1997) and chimpanzees (ripe fruit ; Chapman et al. 1995), could indeed be correlated. However, the typical resources of red colobus (flowers, unripe fruit and young leaves ; Wachter et al. 1997) are mostly out of phase.

258

R. Noe$ and R. Bshary

Red colobus–diana associations

The results of our playback experiments provide the most convincing argument for a direct causal arrow pointing from chimpanzee hunting activity to association behaviour. The monkeys act directly upon information about the presence of chimpanzees in their immediate vicinity. Even during the hunting season the risk of being attacked is relatively low for each single colobus group. In contrast to crowned eagles and leopards, chimpanzees often announce their presence. The monkeys can potentially estimate the momentary risk of predation by one single clue : the hooting and drumming of male chimpanzees in their immediate proximity. On the basis of that information they should be able to trace fluctuations in predation risk in both space and time. Association rates were considerably lower during the hunting seasons in 1993 and 1994 than in 1991 and 1992 for the combination Bad 1¬Dia 1. As the two neighbouring combinations were out of phase in this respect, it is not plausible that a resource-related factor played a role. The unhabituated chimpanzees ranging in our study area avoided us. The monkeys seemingly adjusted to the lower predation pressure after some time lag. This explanation would be in line with the idea that associations are used as a flexible antipredation strategy. A relaxation of the foraging constraints during the season ‘ rain 2 ’ in 1993 and 1994 explains why the association rates were higher than in the corresponding seasons in previous years. (cf. Wachter et al. 1997). During those two seasons one of the few tree species on which the red colobus and diana feed simultaneously provided abundant fruit. The colobus ate the unripe fruit, the diana the ripe fruit. A comparison with a study conducted on Tiwai Island, Sierra Leone, provides further support of the antipredation hypothesis. The monkey community is similar to Taı$ , but the predation is virtually absent (Holenweg et al. 1996 ; Oates, personal communication). The association rates between red colobus and diana on Tiwai Island were not higher than expected under the assumption of chance encounters (Whitesides 1989 ; Oates & Whitesides 1990). 5. C O N C L U S I O N

Predation by chimpanzees largely explains why red colobus monkeys in the Taı$ National Park frequently seek associations with diana monkeys. This suggests that a sufficiently high predation pressure can be a major factor in the evolution of associations between groups belonging to different species. Assuming that mixed-species groups are formed because the limits of monospecific group size are reached, our results also indirectly support the predation hypothesis for monospecific groups. The Ministe' re d’Enseignement Supe! rieur et Recherche Scientifique and the Ministe' re d’Agriculture et Ressources Animales of Co# te d’Ivoire gave permission to work in the Park. We thank the IET, the CSRS, B. Sluijter, and all field assistants and students for their cooperation and support. The comments of P. Hammerstein, H. Hofer, R. Susman, W. Wickler and three anonymous referees helped to improve the manuscript. L. Williams corrected our English. Financial Proc. R. Soc. Lond. B (1997)

support : Swiss NSF, MPIV Seewiesen, Dobberke Foundation and Lucy Burger Foundation. We dedicate this paper to W. Wickler on the occasion of his 65th birthday.

REFERENCES Alexander, R. D. 1974 The evolution of social behavior. A. ReŠ. Ecol. S’st. 4, 325–383. Boesch, C. 1994 a Chimpanzees–red colobus : a predatorprey system. Anim. BehaŠ. 47, 1135–1148. Boesch, C. 1994 b Cooperative hunting in wild chimpanzees. Anim. BehaŠ. 48, 653–667. Boesch, C. & Boesch, H. 1989 Hunting behavior of wild chimpanzees in the Tai National Park. Am. J. ph’s. Anthrop. 78, 547–573. Bortz, J., Lienert, G. A. & Boehnke, K. 1990 Verteilungsfreie Methoden in der Biostatistik. Berlin : Springer. Bshary, R. 1995 Rote Stummelaffen (Colobus badius) und Dianameerkat“en (Cercopithecus diana) im TaıX Nationalpark, ElfenbeinkuX ste : Warum asso“ieX ren sie ? Ph.D. thesis, University of Mu$ nchen. Bshary, R. & Noe$ , R. 1997 Red colobus and diana monkeys provide mutual protection against predators. Anim. BehaŠ. (Submitted.) Chapman, C. A. & Chapman, L. J. 1996 Mixed-species primate groups in the Kibale Forest : ecological constraints on association. Int. J. Primatol. 17, 31–50. Chapman, C. A., Wrangham, R. W. & Chapman, L. J. 1995 Ecological constraints on group size : an analysis of spider monkey and chimpanzee subgroups. BehaŠ. Ecol. Sociobiol. 36, 59–70. Cheney, D. L. & Wrangham, R. W. 1987 Predation. In Primate societies (ed. B. B. Smuts, D. L. Cheney, R. M. Seyfarth, R. W. Wrangham & T. T. Struhsaker), pp. 227–239. Chicago : University of Chicago Press. Conover, W. J. 1980 Practical nonparametric statistics. 2nd edn. New York : John Wiley & Sons. Cords, M. 1987 Mixed-species association of Cercopithecus monke’s in the Kagamega Forest, Ken’a. Berkeley : University of California Press. Cords, M. 1990 Vigilance and mixed-species association of some East African forest monkeys. BehaŠ. Ecol. Sociobiol. 26, 297–300. Doran, D. M. 1989 Chimpan“ee and p’gm’ chimpan“ee positional behaŠior, the influence of enŠironment, bod’ si“e, morpholog’, and ontogen’ on locomotion and posture. Ann Arbor : UMI Dissertation Services. Dunbar, R. I. M. 1988 Primate social s’stems. London : Croom Helm. FitzGibbon, C. D. 1990 Mixed-species grouping in Thompson’s and Grant’s gazelles : the antipredator benefits. Anim. BehaŠ. 39, 1116–1126. Garber, P. A. 1988 Diet, foraging patterns, and resource defense in a mixed species troop of Sanguinus m’stax and Sanguinus fuscicollis in Amazonian Peru. BehaŠiour 105, 18–34. Gautier-Hion, A., ”uris, R. & Gautier, J.-P. 1983 Monospecific vs poly-specific life : a comparative study of foraging and antipredatory tactics in a community of Cercopithecus monkeys. BehaŠ. Ecol. Sociobiol. 12, 325–335. Hamilton, W. D. 1971 Geometry for the selfish herd. J. theor. Biol. 7, 295–311. Holenweg, A. K., Noe$ , R. & Schabel, M. 1996 Waser’s gas model applied to associations between red colobus and diana monkeys in the Taı$ National Park, Ivory Coast. Folia primatol. 67, 125–136. Ho$ ner, O. P., Leumann, L. & Noe$ , R. 1997 Dyadic associations of red colobus (Colobus badius) and diana

Red colobus–diana associations monkey (Cercopithecus diana) groups in the Taı$ National Park, Ivory Coast. Primates. (Submitted.) Inman, A. J. & Krebs, J. 1987 Predation and group living. Trends Ecol. EŠol. 2, 31–32. Isbell, L. A. 1994 Predation in primates : ecological patterns and evolutionary consequences. EŠol. Anthrop. 3, 61–71. Janson, C. H. 1992 Evolutionary ecology of primate social structure. In EŠolutionar’ ecolog’ and human behaŠior (ed. E. A. Smith & B. Winterhalder), pp. 95–130. New York : Aldine de Gruijter. McGraw, W. S. 1996 Cercopithecid locomotion, support use and support availability in the Taı$ Forest, Ivory Coast. Am. J. ph’s. Anthropol 100, 507–522. Oates J. F. & Whitesides G. H. 1990 Association between olive colobus (Procolobus Šerus), diana guenons (Cercopithecus diana), and other forest monkeys in Sierra Leone. Am. J. Primatol. 21, 129–146. Peres, C. A. 1992 a Prey-capture benefits in a mixed-species group of Amazonia tamarins, Sanguinus fuscicollis and S. m’stax. BehaŠ. Ecol. Sociobiol. 31, 339–347. Peres, C. A. 1992 b Consequences of joint-territoriality in a mixed-species group of tamarin monkeys. BehaŠiour 123, 220–246. Pulliam, H. R. & Caraco, T. 1984 Living in groups : is there an optimal group size ? In BehaŠioural ecolog’. An eŠolutionar’ approach. 2nd edn (ed. J. R. Krebs & N. B. Davies), pp. 122–147. Oxford : Blackwell. Sokal, R. R. & Rohlf, F. J. 1995 Biometr’. Third edn. New York : Freeman. Stanford, C. B. 1995 The influence of chimpanzee predation on group size and anti-predator behviour in red colobus monkeys. Anim. BehaŠ. 49, 577–587. Stanford, C. B., Wallis, J., Matama, H. & Goodall, J. 1994 Patterns of predation by chimpanzees on red colobus monkeys in Gombe National Park, 1982–1991. Am. J. ph’s. Anthropol. 94, 213–228. Struhsaker, T. T. 1981 Poly-specific associations among tropical rain-forest primates. Z. Tierps’chol. 57, 268–304. Struhsaker, T. T. & Leakey, M. 1990 Prey selectivity by crowned hawk-eagles on monkeys in the Kibale Forest, Uganda. BehaŠ. Ecol. Sociobiol. 26, 435–443. Terborgh, J. 1983 FiŠe NeW World primates. A stud’ in comparatiŠe ecolog’. Princeton : Princeton University Press. Terborgh, J. & Janson, C. H. 1986 The socioecology of primate groups. A. ReŠ. Ecol. S’st. 17, 111–135.

Proc. R. Soc. Lond. B (1997)

Ronald Noe$ and Redouan Bshary

259

Turner, G. F. & Pitcher, T. J. 1986 Attack abatement : a model for group protection by combined avoidance and dilution. Am. Nat. 128, 228–240. van Hooff, J. A. R. A. M. & van Schaik, J. P. 1992 Cooperation in competition : the ecology of primate bonds. In Coalitions and alliances in humans and other animals (ed. A. H. Harcourt & F. B.M. de Waal), pp. 357–389. Oxford University Press. van Schaik, C. P. 1983 Why are diurnal primates living in groups ? BehaŠiour 87, 120–144. van Schaik, C. P. & van Hooff, J. A.R. A.M. 1983 On the ultimate causes of primate social systems. BehaŠiour 85, 91–117. van Schaik, C. P. & Ho$ rstermann, M. 1994 Predation risk and the number of adult males in a primate group : a comparative test. BehaŠ. Ecol. Sociobiol. 35, 261–272. van Schaik, C. P. & van Noordwijk, M. A. 1985 Evolutionary effect of the absence of felids on the social organization of the macaques on the island Simeulue. Folia primat. 44, 138–147. van Schaik, C. P., van Noordwijk, M. A., Warsono, B. & Sutriono, E. 1983 Party size and early detection of predators in Sumatran forest primates. Primates 24, 211–223. Wachter, B., Schabel, M. & Noe$ , R. 1997 Diet overlap and poly-specific associations of red colobus and diana monkeys in the Taı$ National Park, Ivory Coast. Etholog’. (In the press.) Waser, P. M. 1982 Primate poly-specific associations : do they occur by chance ? Anim. BehaŠ. 30, 1–8. Waser, P. M. 1984 ‘ Chance ’ and mixed-species associations. BehaŠ. Ecol. Sociobiol. 15, 197–202. Whitesides, G. H. 1989 Interspecific associations of Diana monkeys, Cercopithecus diana, in Sierra Leone, West Africa, biological significance or chance ? Anim. BehaŠ. 37, 760–776. Williams, G. C. 1966 Adaptation and natural selection. Princeton : Princeton University Press. Zuberbu$ hler, K., Noe$ , R. & Seyfarth, R. M. 1997 Diana monkey loud calls : messages for conspecifics and predators. Anim. BehaŠ. (In the press.)

ReceiŠed 27 September 1996 ; accepted 29 October 1996