Association between olive colobus (Procolobus verus) Diana guenons (Cercopithecus diana) and other forest monkeys in Sierra Leone.код для вставкиСкачать
American Journal of Primatology 21:129-146 (1990) Association Between Olive Colobus (Procolobus verus), Diana Guenons (Cercopithecus diana), and Other Forest Monkeys in Sierra Leone JOHN F. OATES' AND GEORGE H. WHITESIDES' 'Departmenf of AnthropoloRy, Hunter College and fhe Graduate School, CUNY, New York and 'Department of Biology, Unwersrty of Mtami, Coral Gables, Florula Close association between olive colobus (Procolobus uerus) and other monkeys (especially Cercopithecus species) has been observed throughout the range of P. uerus in the forest zone of West Africa. To investigate the basis of this association, we made new observations in Sierra Leone, concentrated a t Tiwai Island. We obtained data on the association patterns of monkeys over a large area of the island from line-transect samples, and studied association behavior via long-term observational sampling of habituated groups of olive colobus and Diana monkeys (C. dmna). During transect sampling, olive colobus always were seen less than 50 m from monkeys of other species, especially Cercopithecus. In studies of habituated groups, we found that one group of olive colobus associated closely with a larger group of Diana monkeys for more than 3 years. Members of the two groups were within 50 m of each other on over 80% of scan samples; the two groups shared the same range, but foraged in different parts of the canopy and had little dietary overlap; association was maintained by the behavior of the colobus. The olive colobus was the only Tiwai monkey species seen less than 50 m from members of a second Dlana study group more frequently than expected, although in this case the two species were associated during less than 12% of samples. In both cases, we detected month-to-month variation in association frequency. We suggest that olive colobus has a strong tendency to associate with other monkeys a s part of a n evolved strategy; that observed association patterns depend on the ranging habits and group dispersion patterns of the other species in the area; and that this strategy evolved because it reduces predation risk for a small-bodied monkey that forages in small, dispersed groups. Key words: interspecific association, Promlobus verus, Cereopithecus diaria, antipredator strategy INTRODUCTION Nearly all of the best-documented instances of interspecific associatian among primates involve association between members of the same genus, particularly Reeeived for publication December 7, 1989; revision accepted March 6, 1990. Address reprint reqwsts to John F Oates, Department of Anthropology, HunterCollege, 695 Park Ave., New York,NY 10021.Thecurrent address of George H. "hidesides is: Biology Departmeat, GTCC, P.O. Box 309, Jarnestown, NC 27282. 1990 Wiley-Liss, Inc. 130 I Oates and Whitesides Cercopithecus in Africa [e.g., Cords, 1987; Gautier-Hion, 19881 and Saguinus in South America [e.g., Terborgh, 1983; Garber, 19881. Functional explanations for such associations between species which have overlapping diets must tease apart potential costs and benefits of association related to foraging from those related to other factors such as predation risk; such distinctions often are difficult or impossible to make. Some cases of frequent association between members of different genera are known, but these generally involve species with considerable dietary overlap [e.g., Saimiri and Cebus in South America: Klein & Klein, 1973; Terborgh, 19831. Here we document and discuss patterns of association between olive colobus monkeys (Procolobus verus) and guenons (Cercopithecus spp.) in West African forests. These associations are of interest for two main reasons: 1)they involve species with very little dietary overlap, and 2) they suggest that association with other species is a n essential part of the olive colobus adaptive strategy. Over 30 years ago, Booth  noted that the olive colobus often was found feeding in close proximity to Cercopithecus monkeys; he suggested that this was a protective adaptation on the part of the colobus: The more mobile, inquisitive frugivorous guenons, which must move from branch to branch frequently in search of their food, are more likely to spot a n enemy than is the olive colobus, which can sit on a leafy bough and consume several ounces of foliage without shifting positions at all. Moreover, the resonant warning cries of the guenons are observably understood and acted upon by the olive colobus. Booth, however, did not study habituated groups of monkeys, nor did he provide detailed, quantitative statements about diets or patterns of association. Booth suggested that dogs, humans, and crowned hawk-eagles (Stephanoaetus coronatus) might be “enemies” to other species of forest monkey, but he did not provide any specific evidence of predation or attempted predation on olive colobus. In this paper we examine Booth’s ideas in the light of evidence we collected in the course of both line-transect surveys and long-term observations of habituated monkey groups at Tiwai Island in southern Sierra Leone. We pay special attention to a stable association which we observed between one group of olive colobus and one group of Diana monkeys (Cercopithecus diana diana); these two groups lived in more or less permanent association for at least 3 years. THE OLIVE COLOBUS MONKEY The olive colobus is an unusual monkey in several respects. It is the smallest of all colobine monkeys (adult males average 4.7 kg, adult females 4.2 kg) [Oates, 19881, and the most drably colored. It has the most reduced thumb of any colobine. Infants are carried in the mother’s mouth and, at our study site at least, do not cling to her trunk. The species has a peculiar geographical distribution pattern, not found in any other primate; it occurs only in coastal West Africa, in the moist forest zone from southern Sierra Leone to eastern Nigeria; it crosses two important zoogeographical boundaries (the Dahomey Gap and the Niger River), but does not occur east of the Cross River [Oates, 19811. Wherever the olive colobus has been observed in the wild, it has almost always been seen in close proximity to monkeys of other species, especially guenons. Booth [19571 noted that in Ghana he often saw it feeding in the company of Cercopithecus campbelli, C . mona, and C. petaurista. Rucks [19761 often saw olive colobus in groups with C. diana, C. petaurista, and C. campbelli in the Bia National Park, Ghana. During surveys in the Tai National Park (C6te d’Ivoire), Galat-Luong and Galat  detected olive colobus on 24 occasions; in 22 cases these monkeys were Olive Colobus Associations I 131 in close association with other species, most frequently C . diana, C . petaurista, and C . campbelli. At a site in Benue State, Nigeria, J.F.O. saw olive colobus travelling in a mixed group with Cercopithecus mona and C . nictitans [Oates, 19821. In the course of surveys in southern Sierra Leone in 1979,1980, and 1985, we encountered olive colobus eight times; in each case we observed Cercopithecus species within 50 m of the colobus. STUDY SITE Tiwai is the largest (12 km2) and northernmost (7” 33’N, 11”21’W) in a series of partly forested islands in the Moa River. It has sandy soils and a seasonally wet climate; annual rainfall averaged 2,708 mm in 1983-86, with most rain falling in May-November. About 60% of the island is covered by old-growth secondary forest, in which leguminous trees are abundant. In addition to the olive colobus and Diana monkey, other anthropoid primates inhabiting the old forest are the red colobus (Procolobus badius), the black-and-white colobus (Colobus polykomos), the sooty mangabey (Cercocebus atys), the spot-nosed monkey (Cercopithecus petaurista), Campbell’s monkey (Cercopithecus campbelli), and the chimpanzee (Pan troglodytes). Some monkeys were shot by hunters on the island in years prior to 1982, but since that date Tiwai has been protected from hunting. Further details about the site are given in Whitesides et al. , Whitesides , and Oates et al. [19901. METHODS We studied primates on Tiwai by using a set of trails cleared through the forest undergrowth on a 400 m grid. In two areas on either side of the island (East and West Study Areas) this 400 m grid system was subdivided by trails at 50 m intervals. Behavioral and botanical sampling were concentrated in these areas. We collected the observations presented in this paper in two chief ways: 1) by the transect sampling of primate populations in the old forest (this involved repeatedly walking along a set portion of the 400 m grid system), and 2) by systematically sampling the behavior of habituated groups of olive colobus and Diana monkeys inhabiting the two study areas. G.H.W. made twenty-eight 6 km transect samples between 4 October 1982 and 21 January 1984. Anthropoid primates encountered during transect samples were scored as “clusters.” Primates were regarded as belonging to the same cluster if, regardless of species, they were separated by no more than 50 m from any other detected primate. Details of the technique are given in Whitesides et al. [19881. We habituated two groups of Diana monkeys and one group of olive colobus monkeys to tolerate our close proximity. The olive colobus group (“PF”) inhabited the East Study Area, and shared its range with one of the Diana groups (group “E”);habituation of these groups began in February 1983. The second Diana group (“W”) inhabited the West Study Area, and its habituation began in November 1982. We sampled the behavior of these groups systematically while following them from dawn to dusk for periods of 3 consecutive days. G.H.W. sampled group W during two 3-day periods each month from March 1983 through June 1984, and group E for one 3-day period each month from July 1983 through June 1984. J.F.O. sampled group PF during two 3-day periods in July 1983, January 1984, and October 1985, and for one 3-day period in June 1983, August 1983, June 1984, February 1985, and March 1985. We simultaneously sampled olive colobus group PF and Diana group E on January 14-16, 1984, and June 12-14, 1984. At 20-min intervals, we scan-sampled the behavior of individuals within a study group. Scan samples of the olive colobus group lasted 5 min; a t the end of this 132 I Oaks and Whitesides TABLE I. Summary of Results of Monkey Sighting6 From the Twenty-Eight6km Transect Samples Conducted Between October 1982 and January 1984 on Tiwai Island Species -- Cercopitkus campbelli Cercopithecus diana Cercopithecus petuurista Cercmebus atys Colobus polykomos Procolobus hadius Procolohus uerus Total No. of groups“ No. of No. in 5 ’ in othersb association‘ association ------- S2 7 19 36.5 65 12 31 47.1 62 15 26 41.9 19 4 a 42.1 96 15 26 27.1 38 10 20 52.6 17 0 17 100.0 349 63 5 i g h t i n g s consisting of more than one conspecific individual, or if only one individual, then confirmation (usually via vocalizationl of the presence of additional conspecifics. bSightings consisting of only one individual of a species without confirmation of additional conspecific individuals. ‘Because the potential advantages and disadvantages of interspecific association may be very different for lone conspecific individuals compared with yocia1 groups, we report here only associations among social groups. period J.F.O. noted whether, during the sample, individuals of any other anthropoid species had been detected within 50 m of any member of the olive colobus study group. Scan samples of the Diana groups lasted 7 min, or until five individuals were sampled, whichever came first. Olive colobus feeding also was sampled by the “frequency” or “event” method IStruhsaker, 1975; Oates, 19881. For further details of scan sampling see Oates 119881 and Whitesides 119891. At 2 min before and 8 min after each scan sample (i.e., at 10-min intervals) G.H.W. noted t h e presence of any individual heterospecific anthropoids within 50 m of any individual member of a Diana study group, as well as the 50 x 50 m grid cell occupied by members of the study group. RESULTS Transect S a m p l e s During t h e course of 28 transect samples, 307 clusters of monkeys were observed. These clusters contained only a single observed individual or species in 230 cases; 56 clusters contained two observed species; four contained four observed species; and one cluster contained six observed species. Olive colobus monkeys were observed in 17 of the clusters (5.56). In each of these 17 clusters the olive colobus were within 50 m of some member of at least one vther species; in ten clusters one additional species was observed; in four clusters two other species were observed; in two clusters three other species were observed; and in one cluster five additional species were observed. During t h e transect samples, olive colobus was the only species never observed alone (Table I). Olive colobus were observed in association with some species, especially Cercopithecus sp., more frequently than with others (Table 11). These patterns of association differ from expectations based on the relative frequency with which each species was Olive Colobus Associations / 133 TABLE 11. Observed and Expected Frequencies of Association Between Procoloblls verus and Other Monkey Species During Transect Samples on Tiwai Island* Species Cercopithecus campbelli Cercopithecus diana Cercopithecus petauristu Cercocebus atys Colobus polykomos Procolobus badius Total Observed frequency Expected frequency 6 4.2 9 5.3 6 5.0 1 1.6 4 7.8 1 3.1 27 27.0 *Expected frequencies based on relative frequency of sighting of groups of each species during the transect samples (see Table 1). Chi-squared test for observed vs. expected frequency: X2 = 10.499, df = 5,0.1zP>0.05. encountered on transects, although the difference is not quite significant (chisquared one-sample test; X2 = 10.499, 0.1 2 P > 0.05, df = 5). The olive colobus was the first species seen in only two of the 17 clusters in which it was observed. This frequency was significantly less than expected (binomial test: P = 0.0005, N = 17; mean probability that olive colobus first species seen = 0.4412, given the assumptions that: 1)all species were equally likely to be the first seen, and that 2) the observed distribution of cluster sizes seen to contain olive colobus was an accurate reflection of the true distribution of cluster sizes containing olive colobus). Observations of Habituated Groups Group composition and home range size. During the 11three-day samples on the PF group of olive colobus between June 1983 and October 1985, the group contained an average of 8.4 individuals (range 5-10) including an average of 1.4 infants (range 0-2). The same two adult males were present throughout, but the number of fully adult females declined from five to one during this period. Because the samples on both Diana study groups spanned only 12 months, fewer changes in group composition occurred than for the olive colobus study group. Both Diana study groups contained one fully adult male and seven adult females throughout the 12-month period reported here, although the identity of the resident adult male changed in both study groups; differences in size between the two groups resulted entirely from differences in younger age classes. In July 1983, Diana group E contained 15juveniles and four infants, in addition to the adults, giving a total of 27 group members; by June 1984, three juveniles had left the group, and the four infants had become juveniles, producing a total of 24 group members. Diana group W contained 21 group members in July 1983, including six juveniles and seven infants; by June 1984, the group had lost two juveniles, and the seven infants had become juveniles, so that the group totaled 19 members. No infants were born in either group during this 12-month period. 134 I Oates and Whitesides PF group members were recorded in 113 cells of the 50 x 50 m trail grid (28.25 ha), during the 11three-day samples. Because some of these cells bordered the Moa River, and consequently were not complete 50 x 50 m cells, the total land area in which the group was seen totaled about 26 ha. During 12 three-day samples, Diana group E entered 101 full 50 x 50 m grid cells (25.25 h a in area) and 22 of the partial cells bordering the river (3.25 h a in area), for a total area of 28.5 ha. This area was almost identical in both size and location to that used by the PF group of olive colobus which frequently associated with group E. During the 24 three-day samples collected on Diana group W, this group entered 163 cells 50 x 50 m (40.75 ha). If only the first 3-day sample from each month is considered (i.e., the same number of samples as for group E), group W entered 149 grid cells (37.25 ha). Association frequencies. We observed members of the olive colobus PF group and Diana group E in close proximity to each other on most days and during most scan samples. For instance, during 838 scans of PF group in which some olive colobus were detected, J.F.O. observed members of group E less than 50 m from members of PF group on 692 occasions (82.6% of the sample). Likewise, during 2,647 scans (10-min interval samples) of group E, G.H.W. observed members of PF group within 50 m on 2,283 occasions (86.2% of samples). Both of these observed association frequencies represent lower bounds. Scan sample data taken on the olive colobus PF group probably underestimated the true association frequency between that group and Diana group E for two reasons: 1) while a sampled individual may have been greater than 50 m from any Diana group E member, other (undetected) PF group members may have been nearer than 50 m to a group E individual; and 2) a n undetected member of group E may have been less than 50 m from a n olive colobus. The samples taken at 10-min intervals on Diana group E probably also underestimated the true association frequency with the PF group for two similar reasons: 1)the exact location of all group E members could not be determined simultaneously, and 2) the exact location of all PF group members frequently was not known a t the time of the sample; both these factors would cause some degree of underestimation of association frequencies. Diana group W did not show such close association with a particular group of another species as did group E. However, olive colobus were seen within 50 m of members of group W on 11.8%of 5309 samples taken at 10-min intervals. Olive colobus was the only monkey species seen significantly more frequently in association with group W than predicted by a computer simulation taking account of each species’ density and group spread [Whitesides, 19891. The olive colobus seen in association with group W were judged to belong to at least two different groups. Behavior in association. The association between the olive colobus PF group and Diana group E apparently was maintained largely by the behavior of the colobus. The Diana group used a varied foraging pattern-sometimes making slow, steady, amoeboid-like progression, foraging as it travelled; sometimes foraging for several hours within a limited area, and then making relatively rapid linear progressions from one site to another with little foraging along the way. When the Diana group was static or foraging slowly, the olive colobus either rested or fed, choosing a site within or close to the area over which the Diana group was spread. When the Diana group moved to a new area, the olive colobus usually followed, leaving their resting or feeding site as a Diana progression got under way, and mixing themselves with the Diana group. During linear progressions by Diana, olive colobus individuals usually were interspersed with the last one-third of the Diana group. Overt agonistic and affilitative interactions between the two species were seen very rarely. In 833 hours with PF group, J.F.O. saw only one obvious agonism, Olive Colobus Associations / 135 when a n adult female Diana with a n infant lunged toward a n approaching olive colobus, which backed away; G.H.W. saw no agonistic interaction in over 475 hours of observation of group E. Occasionally we saw play between juvenile olive colobus and juvenile Dianas, but members of the two groups appeared to ignore most of each other’s social and sexual behavior. Overt interactions between olive colobus and other monkey species also were rare. On one occasion J.F.O. saw a solitary male spot-nosed monkey (Cercopithecus petaurista) approach and briefly groom one of the adult males in PF group. On another occasion one olive colobus (possibly a juvenile) was seen to run off quickly a t the apparent approach of a n adult red colobus (Procolobus badius). Territorial interactions. During the 36 sample days on Diana group E, G.H.W. observed a t least ten major territorial encounters between group E and a Diana group occupying an area immediately to the north. Each of these encounters occurred on a different day and they were spread over seven different 3-day sample periods. These intergroup interactions included loud calls by the adult male of each group, numerous agonistic vocalizations by most or all group members, and shortduration chases (usually less than 20 m) of members of one group by members of the other group. During five of the ten Diana intergroup encounters (including five of the seven 3-day sample periods), vocalizations were heard both from members of the olive colobus PF group associated with Diana group E, and from other olive colobus apparently associated with the rival Diana group. Many of the olive colobus calls were given in apparent response to calls from the opposite group of olive colobus, resulting in bouts of countercalling. The olive colobus call typically associated with these encounters was the “laughing” call described by Hill and Booth [19571; we heard this call only from adult males. During the 33 sample days on olive colobus P F group, J.F.O. noted ten bouts of loud calling from PF group adult males that were apparently elicited by Diana territorial interactions andlor by the detection of other olive colobus in areas close to the PF home range. (In this case a “bout” refers to calls produced by olive colobus while they remained in a limited part of their range, usually close to the periphery, and while interactions proceeded with other monkey groups which also remained in approximately the same location; the longest of these bouts lasted for just under 3 hours, within which the longest separation of two laughing calls was 29 rnin.) Two of these ten bouts were included in the ten Diana territorial encounters witnessed by G.H.W. In three of the bouts, an adult male olive colobus in PF group was seen sitting in the middle canopy, looking intently in the direction of the opposite group. However, we did not observe more overtly aggressive acts between olive colobus groups. Our assessment of the dynamics of these mixed Diana-olive colobus encounters is that they were initiated largely by different groups of Diana monkeys moving into proximity with one another. Such movement often brought different olive colobus groups into proximity, resulting in adult male countercalling by the colobus. We have no direct evidence that an olive colobus group was “using” Dianas as assistants in territorial defence, or that the outcome of a Diana intergroup encounter was influenced by the behavior of the olive colobus. Responses to calls of other species. Loud calls by olive colobus frequently were clumped temporally with those of one or more other monkey species. On 33 sample days, 360 olive colobus loud calls were noted, clustered in 125 sequences (where one sequence includes individual laughing calls separated by no more than 5 min from other such calls). Of 85 sequences not apparently connected to intergroup encounters, 40 (47.1%) followed within 5 min of the call of another monkey species; in 17 of these cases the first monkey to call was Cercopithecus campbelli, 136 i Oates and Whitesides % Flowers (N = 21) E %Fruit(N=7) 1.8% % Seeds (N = 31) % Pod Exudate (N = 0 ) % Leaf Parts (N = 166) E 0% % Others (N= 0) % Insects (N= 0) A % Flowers (N = 415) % Fruit (N = 528) % Seeds (N = 171) a % Pod Exudate (N= 303) 21.3% El % Leaf Parts (N= 274) I % Others (N = 27) 1.1 S % Insects (N = 764) €3 12.2% Figure l.a,b. and in 16 cases the first caller was C . diana. Only in one instance was a n olive colobus call first in a sequence of loud calling by several monkey species. Loud calls by Diana monkeys also were frequently associated temporally with Olive Colobus Associations / 137 0 % Flowers (N = 167) %Fdt(N=342) % Seeds (N= 55) %PodExudate(N=90) El C1.8% % Leaf Parts (N= 151) % Others (N = 10) H 14 % Insects (N= 265) C 5.1% Fig. 1. A Proportion of time spent feeding on each category for the PF group of colobus (N = 225). B Proportion of time spent feeding on each food category for Diana group W (N = 2,482). C: Proportion of time spent feeding on each food category for Diana group E (N = 1,080). those of other primate species. Of 234 Diana loud calls noted by G.H.W. in the East study area, 142 (60.7%) occurred in close temporal relation to loud calls given by other primates, most frequently other Diana monkey loud calls. Diana monkeys frequently gave distinctive, high-pitched alarm calls in response to the alarm calls of other animals, such as squirrels, duiker antelopes, and birds. Olive colobus also appeared to respond to the alarm calls of other species, including Diana monkeys, by increased vigilance behavior and frequently by shifting to less conspicuous or less exposed positions. The functions of a variety of low-amplitude olive colobus calls were not elucidated satisfactorily in this study, and we could not recognize a distinctive alarm call produced by the olive colobus. Nor did we detect that olive colobus and Diana monkeys had special responses to each other’s calls, different in quality from their responses to the calls of other species. Duration of association. Our own observations combined with those of other researchers studying primate behavior on Tiwai (particularly G.L. Dasilva, A.G. Davies, and C.M. Hill) showed that Diana group E was in existence (and using approximately the same range) from at least February 1983 through December 1988. Throughout this period, olive colobus monkeys frequently were seen less than 50 m from the Diana monkeys. Recognizable individual olive colobus (serving to identify the animals as members of the PF group) were first noted (by J.F.O.) moving with the Diana group in March 1983 and last recorded in January 1987; before and after these dates the same individuals may have associated with the Diana group, but our observations do not include those periods. Diets. The diets of olive colobus and Diana monkeys on Tiwai are very different. The foods most frequently recorded in sampling the diet of olive colobus PF 138 I Oates and Whitesides group were young leaves, seeds, and mature leaf petioles, while the most frequently sampled foods for both Diana study groups were ripe fruits, insects, and flowers (Fig. 1). If specific food items are considered (i.e., a food species plus the part of that species consumed), there was little overlap between the diets of olive colobus PF group and Diana group E, although they used almost exactly the same area of forest. The top ten items in the diet of each group are shown in Table I11 with no overlap in items between the two species. (Note that the olive colobus diet data presented here are from scan samples, while the data in Oates are from event sampling.) Activity heights. Figure 2 compares the distribution, based on scan samples, of olive colobus and Diana activities across heights in the forest canopy. On average, the olive colobus used lower heights in the canopy than either of the Diana study groups, despite frequent travel with Diana E group. In 38.4% of scan samples on the olive colobus PF group, the monkeys were below a n estimated height of 15 m, compared with 25.2% of samples on Diana group E, and 27.6% of samples on Diana group W. Although 67.7% of olive colobus travel was at or above 15 m, 53.5% of feeding was recorded below this height. The time spent by olive colobus feeding low down in patches of thick growth probably was underestimated; often, the monkeys were not seen clearly when foraging in this fashion, and in those cases they were not included in scan sample records. Diana monkeys spent most of their time in or just below the crowns of upper canopy and emergent trees. The upper canopy of the forest in the East Study Area was less broken than that in the west area, but the East Area had fewer very tall, emergent trees; this difference probably accounts for the greater frequency of group E samples between 15 and 24 m, and the greater frequency of W group samples between 25 and over 34 m. Seasonal variation in association. Variation in the frequency of association between the PF group of olive colobus and Diana group E occurred among sampling periods. Diana monkeys were seen within 50 m of PF group members on less than 80% of PF group scans during dry season samples (January 1984, FebruaryMarch 1985) and in a sample a t the end of October 1985 (late wet season) (Table IV). In samples of Diana group E, olive colobus were detected within 50 m of the Dianas on less than 80% of samples in January 1984 and in August and September 1983 (wet season) (Table IV). Major seasonal changes in the diets of both olive colobus [Oates, 19881 and Diana monkeys (Whitesides, unpublished data) occurred at Tiwai. These changes probably contributed to variations in association frequency. For instance, during the January 14-16,1984, sample period, when we followed olive colobus PF group and Diana group E simultaneously, we observed PF group members spending a n unusual amount of time close to the Moa River bank, feeding on the seeds of the river-fringing tree Plagiosiphon emarginatus. This food item was available for only a limited part of the year, a time when scarcity existed of some of the young-leaf foods favoured by the olive colobus [Oates, 19881. While the olive colobus fed near the river during this January sample, most of the Diana group foraged inland. In this case, the attraction of the colobus to a particular food item apparently overrode their usual attraction to the Dianas. The frequency of association of olive colobus with Diana group W also varied from month to month; olive colobus were seen least frequently with group W in August 1983 and in November 1983-March 1984 (Table IV). Association of olive colobus with other species. On the few occasions when the olive colobus PF group was not associated with Diana group E for several hours, they commonly associated with a group of another species rather than Olive Colobus Associations I 139 TABLE 111. Top Ten Plant Items in the Diet of Procolobus uerus (PF Group) and Cercopithecus dianu (Group E and G r o u p W)on Tiwai Island, Sierra Leone* P. uerus (PF group) (N = 211) Rank Item 1 Terminalia ivorensis (young leaves) Sapium aubrevillei (young leaves) Acacia pennata" (young leaves) Sapium aubreuillei (mature leaf petioles) Lovoa trichilioides (flowersiflorai buds) Plagiosiphon emarginatus (immature seeds) Parkia bicolor (young leaves) Funtumia africana (mature seeds) Myrianthus libericus (young leaves) Pterocarpus santal inoides (young leaves) 2 3 4 5 6 7 8 9 10 C. diana (group E) (N = 1080) % Item C. diana (group W) (N = 2482) Item % 10.9 Uapaca guineensis (mature fruit) 9.5 Landolphia hirsuta" (mature leaf petioles) 5.2 Funtumia africana (immature pod fluid) 5.2 Funtumia africana (nectar) 4.7 Dialium guinense (mature seeds) 4.7 Millettia leonensis" (flowers) 3.3 3.3 2.8 2.4 22.3 Funtumia africana (immature pod fluid) 8.0 Uapaca guineensis (mature fruit) 7.8 Funtumia africana (nectar) 6.9 Millettia leonensis" (flowers) 4.8 Landolphia kirsuta" (mature leaf petioles) 4.4 Combretodendron macrocarpum (immature seeds) Landolphia hirsuta" 3.9 Landolphia hirsuta" (mature fruit) (mature fruit) Daniellia ogea 1.4 Millettia leonensis" (flowers) (immature leaves) Unidentified sp. liana" 1.3 Daniellia ogea (flowers) (flowers) Unidentified sp. liana" 1.3 Ficus sp. 1 (fruit) (mature fruit) % 11.6 10.2 8.1 5.5 4.9 4.8 3.8 2.1 1.5 1.3 - *Only observations obtained during scan sampling used in summary presented here. Sample sizes represent known food items. "Food species is liana rather than tree. travelling alone. Other monkey species were detected within 50 m of PF group members in 63 of 139 five-min sample periods (45.3%)in which no Diana monkeys were detected within 50 m of the olive colobus. In these 63 samples, spot-nosed monkeys (Cercopithecus petaurista) were the other species most frequently detected (30 samples), and black-and-white colobus (Colobus polykomos) were the second most common associate (21 samples). Spot-nosed monkeys typically moved in a quiet, skulking fashion in thick growth and often were detected only by their alarm calls, which they commonly produced if approached closely by humans. When members of olive colobus PF group were seen in association ti.e., within 50 m) with spot-nosed monkeys, the two species frequently appeared to be moving together in a coordinated fashion, with each species apparently playing a similar role in maintaining the association. Our observations in the West Study Area at Tiwai suggested that at least two small groups of olive colobus overlapped the home range of Diana group W. One of these groups frequently was seen in close proximity to spot-nosed monkeys; apparently, the olive colobus associated predominantly with the spot-nosed group, but temporarily followed Diana group W on some occasions when the Dianas were nearby (i.e., just the reverse of the situation involving the PF group of olive colobus and Diana group E). Our strong impression was that olive colobus on Tiwai adopted some of the gross behavioral patterns of other species with which they routinely associated. 140 / Oates and Whitesides % Group W (Diana; N = 10,629) % Group E (Diana;N = 5641) R % PF group (olive colobus; N = 1148) 0 0-4 5-9 10-14 15-19 20-24 25-29 30-34 >34 HEIGHT (m) 1m - 80- 60- - 40- 7% Group W (Diana; N = 10,629) % Group E (Diana; N = 5641) _...e... 7% PF group (olive colobus; N = 1148) m- 59 1014 1519 2024 2529 3034 >34 6 HEIGHT (m) Percentage (A) and cumulative percentage (B) of individuals sampled for each height category for each Olive Cdobus Associations / 141 TABLE W.The Percentage of Samples and Sample Sizes by Month in Which Each Study Group Was Observed Within 50 sn of at Least One Individual of the Indicated Species (the Defkition of Association)* PF group with C. diana DATE June 1983 July 1983 August 1983 September 1983 October 1983 November 1983 December 1983 January 1984 February1984 March 1984 April 1984 May 1984 June 1984 February 1985 March 1985 October 1985 Totals Group E with P. uerus Group W with P. uerus % (N) % (N) 70 (N) 92.1 (89) 95.7 (162)" 83.6 (51) - - 68.8 (154Y 94.0 (83) 63.3 (71) 70.7 (75) 85.3 (143)" 82.6 (838) 93.7 (221) 77.9 (222) 79.6 (220) 94.4 (213) 83.7 (208) 89.3 (214) 70.8 (216) 81.9 (226) 98.2 (226) 82.5 (228) 83.9 (223) 98.7 (230) - 37.2 3.4 40.4 10.1 3.9 3.0 1.4 0.9 0.7 11.6 8.7 20.0 (438)" (444)" (438)" (426)" (431)" (429)" (442)" (447)" (454)" (447)" (459Y (454)" - - 86.2 (2,647) 11.8 (5,309) *For the P F group ofProcoZobusuerus data derived from scan samples; for group E and group W of Cercopithecus diana data derived from 10-min samples (group samples). aIndieates samples combined from two different 3-day sampling periods during the same month. Thus, we observed olive colobus relatively easily when they associated with Diana monkeys, which are conspicuous and spent much of their time in the upper canopy. We observed them with greater difficulty when they associated with the quiet, stealthy, spot-nosed monkeys which spent much time in vine tangles and the thick growth in canopy gaps. Galat and Galat-Luong  also noted this tendency of the olive colobus to adopt behavior patterns of species with which it is associating, leading to its designation of "Singe Magique" by hunters in CBte d'Ivoire. Olive Colobus at Pujehun The short-term observations at Pujehun in 1979 and 1980 suggested that the olive colobus there shared a large part of their range with one group of spot-nosed monkeys and one group of Campbell's monkeys. J.F.O. observed olive colobus associating with these groups for several hours, switching from one group to the other on occasions when the two Cercopithecus groups came together and then diverged. The olive colobus apparently were not moving permanently with either group; however, our observations were very limited in duration and therefore we do not have records of the long-term patterns behavior of these monkeys. On one occasion at Pujehun a n adult Campbell's monkey was seen grooming an adult olive colobus. DISCUSSION Potential advantages for species involved in interspecific associations usually fall within two broad categories: foraging advantages and predator avoidance [Whitesides, 19891.The proposed foraging advantages include: 1) reduction of duplication of effort and regulation of return time to resources, 2) increased detection of re- 142 I Oates and Whitesides sources, 3) availability of food items otherwise unavailable, 4)increased prey capture rate, and 5) increased intraspecific competitive ability. Postulated predator avoidance advantages include: 1) increased detection of predators, 2) increased predator confusion, 3 ) decreased probability of discovery or capture by predators, and 4)increased defense against predators. Potential benefits also could arise from a combination of foraging and predator avoidance advantages (e.g., a decrease in necessary vigilance behavior provided by association, thus allowing greater foraging efficiency). For detailed references to the above hypotheses, see Whitesides . Our observations confirm that olive colobus have a very strong tendency to associate closely with other monkey species, especially arboreal guenons. Whitesides  found that olive colobus was the only species which associated with the two study groups of Diana monkeys more frequently than expected by chance. We suggest that the chief benefit that olive colobus derive from associating with these other species is a reduction in the risk of predation. Due to the lack of dietary overlap between olive colobus and Diana monkeys (and almost certainly the other arboreal guenons as well), most of the proposed foraging-advantage hypotheses can be rejected; those hypotheses requiring substantial dietary overlap between species to be effective include: 1)reduction of duplication of effort and regulation of return time, 2) increased detection of resources, and 3 ) availability of food items otherwise unavailable. In addition, the proposed advantage of increased prey capture rate can be rejected, because we never observed olive colobus foraging for or eating insects [Oates, 19881. Only the final foraging advantage (increased intraspecific competitive ability) cannot be easily rejected; however, due to the ecology and behavior of olive colobus, we also judged this hypothesis highly unlikely. All or nearly all of the proposed predator avoidance advantages appear highly plausible. Olive colobus are small-bodied monkeys that typically live in small (less than ten individuals) social groups which tend to become relatively highly dispersed during foraging. Such dispersion probably is related to the dispersion of young leaves in the liana tangles and gaps in which the species frequently forages [Oates, 19881. For small-bodied animals living in small groups and foraging in thick growth, obvious advantages exist in associating with the larger social groups of species with which they have virtually no dietary overlap, such as the Diana monkeys on Tiwai. Factors additional to a lack of dietary overlap may influence the choice by olive colobus of guenons a s species with which to associate in reducing predation risk. We noticed that a human observer may approach West African colobus monkeys without detection more easily than guenons. Although we do not know the reason for this, one contributing factor might be differences in the evolved functions of visual and neural systems as they relate to foraging behavior (for instance, insects are important in guenon diets). The tendency of olive colobus to associate with other species often results in a n association of several hours in duration (as we witnessed in the West Study Area at Tiwai and at Pujehun, and as the observations of Galat & Galat-Luong [19851 in Tai suggest). However, our observations in the East Study Area a t Tiwai demonstrate that the association can be a more or less permanent phenomenon. The pattern and duration of association adopted by a particular olive colobus group probably depend on the ranging and foraging patterns of other monkey groups using the same area, and on local patterns of food availability. For instance, Galat and Galat-Luong [19851reported home ranges of 29 and 33 h a for groups of six and seven olive colobus monkeys in the Tai forest, very similar to the 26 h a used by PF group on Tiwai. But while Diana group E used a range of Olive Colobus Associations / 143 28.5 ha, a Diana group in the Tai study area had a home range of 93 ha [Galat & Galat-Luong, 19851. Similarly, Diana group W a t Tiwai had a larger range size (40.75 ha) than group E, and had only temporary associations with olive colobus. We argue that, if olive colobus have a choice of other-species groups with which to associate, the best choice is a group of guenons with a home-range similar in size to that needed to provide a year-round food supply for the olive colobus. If the group of guenons has a foraging pattern compatible with olive colobus needs, then a more or less permanent association may result. Such association would give the added advantage of producing familiarity with the subtleties of the other species’ behaviors, especially its reaction to predators. Furthermore, the long-term nature of the association would probably be reinforced by the learning experiences of young animals. Olive colobus may associate more frequently with Diana monkeys than other guenons at Tiwai not only because of a similarity in ranging needs, but also because of the relatively large area over which a Diana monkey group is dispersed while foraging. Members of a single social group of Diana monkeys typically are spread over an area greater than 100 m in diameter, while at times exceeding 200 m in diameter [Whitesides, 19891; members of other social groups of guenons rarely were spread over a n area greater than 50 m in diameter with 20 to 30 m being the more typical group spread (J.F.O., G.H.W., personal observation; R. Kluberdanz, G. Davies, G. DaSilva, S. Green, and L. White, personal communications). Such large group spreads of Diana monkeys have two potential advantages for the olive colobus: 1) a larger array of potential feeding sites covered by the protective “umbrella” of a guenon group, and 2) a more far-reaching early warning system against predators, even when feeding is not in progress. We found interactions between olive colobus and other monkeys (at least Diana monkeys) to be essentially one-way. The olive colobus orient positively to the other species, and are largely ignored in return. The guenons seem to suffer little or no detriment from being followed by the colobus, and possibly gain some benefit from the presence of extra animals that can provide warning of potential predators. A species such as the Diana monkey, which uses a different part of the forest canopy and has a very different diet than the olive colobus, has the greatest likelihood of receiving such benefits. Without incurring costs from increased feeding competition, Diana monkeys may obtain warnings about predators from a layer of the forest which they are not able to monitor closely themselves. The tendency of olive colobus to associate with other monkeys may be part of a genetically based evolved strategy. The behavior has been seen in all observed populations of olive colobus, even those (such as in eastern Nigeria) that have had no recent contact (either behavioral or genetic) with other populations. Other components of such a n evolved strategy probably include the species’ drab coloration, low vocalization rate, and habit of spending much time inactive in thick growth. A small group of such animals is very difficult to detect within a larger group of noisier, more active, more colorful guenons as demonstrated by our data from the transect surveys on the sequence of detection of olive colobus within a n association. What predators might olive colobus avoid by this behavior? Struhsaker [19811 has argued that crowned hawk-eagles (Stephanoaetus coronatus) are a major factor producing interspecific association in African forest monkeys generally. The olive colobus, however, due to its tendency to feed in thick, tangled, low growth, probably is less susceptible to attack by crowned hawk-eagles than are most African forest monkeys. Although smaller forest raptors might pose a threat to young olive 144 / Oates and Whitesides colobus monkeys, perhaps mammalian and not avian predation has been the major selective force in producing the olive colobus strategy. Among potential mammalian predators on olive colobus are chimpanzees and leopards, as well as humans. Until very recent times, most olive colobus populations probably shared their habitat with chimpanzees and leopards (and the immediate ancestor of the olive colobus very probably was sympatric with ancestral forms of these larger mammals). Chimpanzees inhabit Tiwai, and suggestive evidence indicates that leopards also may be present. Chimpanzees, leopards, and olive colobus co-occur in the Tai National Park in CGte d’Ivoire, where the forest ecosystem is similar to that at Tiwai. In Tai, Hoppe-Dominik [19841 studied the predation patterns of both chimpanzees and leopards. In a sample of 215 leopard scats from Tai, the olive colobus was the only one of seven Tai monkey species whose hair was not specifically identified. In their study of the hunting behavior of Tai chimpanzees, Boesch and Boesch [19891 noted that groups of chimpanzees searched intentionally for middle-sized mammals, especially red colobus. The red colobus (Procolobus badius), a close phylogenetic relative of the olive colobus, apparently is a major prey species for chimpanzees in more than one part of Africa [e.g., Gombe, Tanzania: Busse, 1977; Mahale Mountains, Tanzania: Takahata et al., 19841. Boesch and Boesch  recorded two attempts by chimpanzees to hunt olive colobus in Tai, during a 7-year period (both attempts were successful) compared with 110 attempts on (and 63 captures of) red colobus; a t Tai the population density of red colobus was estimated at about 14 times as great as that of olive colobus [Galat & Galat-Luong, 19851. Boesch and Boesch  recorded very few attempts by chimpanzees to hunt Cercopithecus monkeys, noting that C . diana in particular “are very agile and much quicker than the colobus; they are hunted only when very low in the canopy.” Tai chimpanzees search for prey while moving on the ground, but will climb into the canopy to catch their victims. We suggest that the olive colobus (similar in many ways to the red colobus, but with half the body size, living in smaller groups, and feeding nearer the ground) would face a high risk of predation from chimpanzees unless they adopted a special counterstrategy. The evidence from Tai of low predation levels on olive colobus suggests that the species has a relatively effective anti-predator strategy. Although this evidence (along with other evidence we have presented) does not prove that association behavior evolved primarily in response to predation pressure, it is at least consistent with this hypothesis. Clearly, primate interspecific association behavior is unlikely to have a single cause across (or even within) taxa, but in this case association as part of a strategy to lower predation risk seems the most plausible explanation of current evidence. CONCLUSIONS 1. Observations in the rain-forest zone of Sierra Leone confirmed that olive colobus monkeys have a strong tendency to associate with monkeys of other species, especially members of the genus Cercopithecus. 2. One olive colobus group a t Tiwai Island associated closely with a group of Cercopithecus diana for more than 3 years. We studied this association and found that it was maintained by the behavior of the colobus, which largely were ignored by the Diana monkeys. The two species shared the same home range, but concentrated their feeding in different parts of the forest canopy and had little dietary overlap. However, some associations between olive colobus and groups of other species were more temporary (on the order of several hours). Olive Colobus Associations / 145 3. We argue that the particular species with which an olive colobus group associates in any given area, and the duration of this association, are related to the ranging and feeding patterns of other species in that area, to patterns of food availability, and to the distances in the forest canopy over which foraging groups of other species are spread. 4. The olive colobus potentially is very vulnerable to predators because it is a relatively small-bodied monkey that forages in small, dispersed groups. We suggest that the association behavior of the olive colobus is part of a strategy that has evolved because it reduces predation risks. One potentially dangerous predator of olive colobus is the chimpanzee. ACKNOWLEDGMENTS This work was funded by grants from the National Science Foundation (BNS 8120206, 8505702 to J.F. Oates and S.M. Green), the Research Foundation of CUNY, and the New York Zoological Society. For their help in Sierra Leone we are very grateful to: the people of Barri and Koya Chiefdoms (especially Chiefs V.K. Magona and M.M. Kanneh, the people of Kambama and Mapuma, and Dr. S.S. Magona), the staff of Njala University College (in particular Mr. G. Goba, Dr. A. Sesay, and Dr. P.T. White), personnel of the Forestry Division (Freetown), the Rev. Pat Palmer (Pujehun), and Nigel and Rachel Wakeham (Freetown). We thank Marilyn Norconk for inviting us to participate in the symposium at the IPS Congress in Brasilia at which an early version of this paper was presented. We also thank three anonymous reviewers for their critical comments on earlier versions of the manuscript. REFERENCES Boesch, C.; Boesch, H. Hunting behavior of Garber, P.A. Diet, foraging patterns, and resource defense in a mixed species troop of wild chimpanzees in the Tai National Saguinus mystax and Saguinus fuscicollis Park. AMERICAN JOURNAL OF PHYSin Amazonian Peru. BEHAVIOUR 105: ICAL ANTHROPOLOGY 78:547-573, 18-34,1988. 1989. Booth, A.H. Observations on the natural his- Gautier-Hion, A. Polyspecific associations among forest guenons: ecological, betory of the olive colobus monkey, Procolohavioural and evolutionary aspects. PD. bus uerus (van Beneden). PROCEEDINGS 452-476 in A PRIMATE" RADIATION: OF THE ZOOLOGICAL SOCIETY OF EVOLUTIONARY BIOLOGY OF THE LONDON 129:421-430,1957. AFRICAN GUENONS. A. Gautier-Hioc Busse, C. Chimpanzee predation as a possiF. Bourliere, J.-P. Gautier, J. Kingdon, eds. ble factor in the evolution of red colobus Cambridge, Cambridge University Press, monkey social organization. EVOLUTION 1988. 31:907-911, 1977. Cords, M. Mixed-species association of Cer- Hill, W.C.O.; Booth, A.H. Voice and larnyx in African and Asian Colobidae. JOURcopithecus monkeys in the Kakamega ForNAL OF THE BOMBAY NATURAL HISest, Kenya. UNIVERSITY OF CALIFORTORY SOCIETY 54:309-321,1957. NIA PUBLICATIONS IN ZOOLOGY 117: Hoppe-Dominik, B. l h d e du spectre des 1-109, 1987. proies de la panthhre, Panthera pardus, Galat, G.; Galat-Luong, A. La communaute dans le Parc National de Tai en CBte d'de primates diurnes de la forGt de Tai, CBte Ivoire. MAMMALIA 48:477-487, 1984. dIvoire. REVUE DECOLOGIE (TERRE & Klein, L.L.; Klein, D.J. Observations on two VIE) 40~3-32,1985. types of Neotropical primate intertaxa asGalat-Luong, A.; Galat, G. ABONDANCE sociations. AMERICAN JOURNAL OF RELATIVE ET ASSOCIATIONS PLUPHYSICAL ANTHROPOLOGY 38:649RISPECIFIQUES DES PRIMATES DI654,1973. URNES DU PARC NATIONAL DE TAI, COTE DIVOIRE. Adiopodoume, CBte d'- Oates, J.F. Mapping the distribution of West African rain-forest monkeys: issues, methIvoire, 1978. 146 I Oates and Whitesides ods, and preliminary results. ANNALS OF THE NEW YORK ACADEMY OF SCIENCES 37653-63, 1981. Oates, J.F. In search of rare forest primates in Nigerai. ORYX 16:431-436, 1982. Oates, J.F. The diet of the olive colobus monkey, Procolobus uerus, in Sierra Leone. INTERNATIONAL JOURNAL OF PRIMATOLOGY 9:457-478, 1988. Oates, J.F.; Whitesides, G.H.; Davies, A.G.; Waterman, P.G.; Green, S.M.; Dasilva, G.L.; Mole, S. Determinants of variation in tropical forest primate biomass: new evidence from West Africa. ECOLOGY (in press), 1990. Rucks, M. NOTES ON THE PROBLEMS OF PRIMATE CONSERVATION IN BIA NATIONAL PARK. Report to Department of Game and Wildlife, Accra, Ghana, 1976. Struhsaker, T.T. THE RED COLOBUS MONKEY. Chicago, University of Chicago Press, 1975. Struhsaker, T.T. Polyspecifie associations among tropical rain-forest primates. ZEITSCHRIFT FUR TIERPSYCHOLOGIE 57:268-304,1981. Takahata, Y.; Hasegawa, T.; Nishida, T. Chimpanzee predation in the Mahale Mountains from August 1979 to May 1982. INTERNATIONAL JOURNAL OF PRIMATOLOGY 5213-233, 1984. Terborgh, J. FIVE NEW WORLD PRIMATES: A STUDY IN COMPARATIVE ECOLOGY. Princeton, Princeton University Press, 1983. Whitesides, G.H. Interspecific associations of Diana monkeys, Cercopithecus diana, in Sierra Leone, West Africa: biological significance or chance? ANIMAL BEHAVIOUR 37:760-776,1989. Whitesides, G.H.; Oates, J.F.; Green, S.M.; Kluberdanz, R.P. Estimating primate densities from transects in a West African rain forest: a comparison of techniques. JOURNAL OF ANIMAL ECOLOGY 57:345367, 1988.