Diet and habitat preference of Mandrillus sphinx in gabon Implications of foraging strategy.код для вставкиСкачать
American Journal of Primatology 11:9-26 (1986) Diet and Habitat Preference of Mandrillus sphinx in Gabon: Implications of Foraging Strategy SALLY A. LAHM Zoological Society of S a n Diego An ecological study of the mandrill was conducted in northeastern and central Gabon from November 1982 to October 1983. The purpose of the study was to gather basic ecological data on the mandrill as a foundation for a future long-term study of this species. Data were collected by direct observation, by collecting evidence left along fresh trails, and by fecal and stomach content analyses. Fruits constituted the mandrill’s major dietary item, supplemented by various plant parts and numerous insect species. Small vertebrates were also occasionally consumed. Mandrills fed primarily on the forest floor but also climbed trees to obtain food, probably on a more frequent basis than do the Papio species. The majority of mandrill sightings and identified foods were attributed to primary forest, but foraging also occurred in secondary, riparian, and inundated forests. The patchy distribution and seasonal fluctuation of fruiting trees in the rainforests of Gabon may influence mandrill home range usage. The eclectic feeding behavior and large home ranges estimated for mandrill groups suggest that this species may play a n important role in seed dispersal. Interspecific competition for food sources may be mitigated by species’ preferences for different fruit parts. The mandrill is able to utilize foods in both arboreal and terrestrial ecological niches. Key words: mandrill, ecology, seed dispersal, competition INTRODUCTION The behavior and ecology of the mandrill (Mandrillus sphinx) are scarcely known. Short-term studies of the mandrill by Jouventin  and Sabater Pi  provided the only information about this elusive species until the recent publication by Hoshino et a1  of a n extended study of mandrill ecology in Cameroon. Despite these commendable efforts, the socioecology of mandrills is still not known in detail owing to their timidity, wide-ranging habits, and densely vegetated habitats. The mandrill is the largest member of the Cercopithecinae. Adult males attain weights of 30-40 kg, while females are nearly half the size of the males [Dorst & Received October 29, 1985; revision accepted March 11, 1986. Address reprint requests to Sally A. Lahm, Department of Anthropology, New York University, 25 Waverly Place, New York, NY 10003. 0 1986 Alan R. Liss, Inc. 10 / Lahm Dandelot, 19701. To justify the placement of the mandrill in a genus separate from that of Papio, Napier and Napier  and Jolly  have cited a number of morphological characters related to the facilitation of increased branch-walking and fruit-processing as evidence of the mandrill's considerable divergence from the Papio species. The geographical distribution of the mandrill has not been definitely determined. Grubb  proposed that the mandrill occurs in Cameroon south of the Sanaga River, Rio Muni, Gabon, and Republic of the Congo. The southern and eastern limitations of its range are not known, although the Ivindo River is purported to be a barrier to the mandrill's eastward extension in Gabon [Jouventin, 19751. The mandrill is officially listed as a n endangered species owing to hunting pressure and habitat alteration throughout its limited range [Wolfheim, 19841. Study Areas Gabon is situated on the equator (3'N-3'S) on the west central African coast between longitudes 8" and 15" E. It covers a n area of 267,667 km2 and is composed of 15% savannah and 85%tropical rain forest. The climate is characterized by two dry and two rainy seasons annually. Amounts of rainfall vary among different regions, ranging from 1,200 mm annually in the southeast to 3,200 mm along the coast [Hladik, 19781. Field work was conducted in two distinct geographical areas (Fig. 1). The northeastern region near Makokou is characterized by high diversity of vegetation and continuous rainforest. More than 900 species of plants have been identified, the dominant tree species consisting of the subfamilies Mimosoideae and Papilinoideae [Caballe, 1978; Florence & Hladik, 19801. The temperature averages 24"C, with humidity remaining constant near 85%. Rainfall averages 1,700 mm annually [Hladik, 19781. CAMEROUN 1 - ) CONGO -0 . ,............,.............. 200 Km STUWSITES Fig. 1. Locations of study areas in Gabon (reprinted from Primate Conservation by permission, S. Nash). Diet and Habitat Preference of Mandrill / 11 The Reserve de Lope-Okanda in central Gabon comprises 5,000 km2 of savannah, gallery, and hill forests at 200-500 m elevation. Rainfall averages 1,600 mm annually. Although there is a considerable overlap of plant species between the two ecological zones, the floristic composition of the central region is dominated by the families Cesalpiniaceae and Myristicaceae [Caballe, 1978; de Saint Aubin, 19631. METHODS An ecological study of the mandrill was conducted in northeastern and central Gabon from November 1982 to October 1983 based a t the Institut de Recherche en Ecologie Tropicale near Makokou. Data were gathered by direct observation, by collecting evidence left along fresh trails, and by fecal and stomach content analyses. Initial surveys were conducted in several localities in northeastern and central Gabon, while actual field work was done in the vicinity of the Institute and at the Reserve de Lope-Okanda in central Gabon. The study yielded data on mandrill diet and habitat preference and additional information regarding group size, composition, defense, and predation. Mandrills were observed during 14 encounters, ranging from 17 to 210 minutes and averaging 72.5 minutes duration, for a total of 16.91 hours of stationary visual contact. Much additional time was logged while following groups, since the rough terrain, dense vegetation, and animals’ timidity prevented prolonged contact and observation. Groups rarely remained in one location for longer than 10-20 minutes. All collected fecal samples were preserved in 10% formalin solution and labeled according to date and location for future microscopic analysis. It was not possible to attribute fecal samples to individual mandrills. Since frequent and prolonged observation was not possible, additional dietary information was collected by performing stomach content analysis. Although the mandrill is hunted and consumed by the rural population, i t is not frequently encountered on hunting forays. Arboreal monkeys and Cephalophus (duiker) species are the major game sources of protein for native inhabitants. Of the 12 mandrill stomachs and skulls acquired from local hunters, six were attributed to adult males, one to a juvenile male, three to juvenile females, and two to adult females. Assessment of age and gender were provided from cranial and dental measurements of the specimens and through information from native hunters. Stomach contents were dried, weighed, and filtered through several sieves of varying mesh sizes. The remnants of each sieving were sorted into various food categories with a binocular microscope. These included fruits and seeds; invertebrates; bark, leaves, stems, and fiber; earth; and fungi. Fruits and seeds were placed in the same category, since it was not always possible to separate seeds from fruit pulp. Bark, leaves, stems, and fiber were combined, as they appeared frequently but in minute portions. Each food category was weighed separately. Fecal samples were prepared similarly [Gautier-Hion, 19801. Forty-one percent of all dietary items were collected during direct observation, with fecal and stomach content analyses contributing 30% and 29%, respectively. The value of these two methods in primate diet analysis has been discussed [Moreno-Black, 19781. While it is true that fecal and stomach content analyses do not accurately reflect a species’ entire dietary regime but only that which has recently been eaten, this bias is compensated for by variations in the localities and seasons of collections. Since it was not possible to positively identify mandrill groups and since fecal samples and stomach contents were collected in substantially separated areas within both ecological zones, it is thought that the results of the analyses are representative of a number of different groups. The small sample size (63 scats, 12 stomachs) offers limited interpretation but may reflect seasonal availability and 12 I Lahm preference of food items. Some invertebrates, in particular, were accorded only familial taxonomic status owing to the scant remains available. Most insect and animal specimens consisted of fragmented body parts. Not all botanical specimens were identifiable. In some cases, a few seeds were the only evidence of a fruit species’ consumption. RESULTS Plant Foods Sixty-three species of plants and fruits in 27 families were identified as mandrill foods. Of these, the families Annonaceae, Sapotaceae, and Olacaceae were the best represented. Large fruit-bearing plants predominated as food sources: 72% were trees, 9%were lianas, and 11% were undetermined. Other plant parts eaten included leaves, bark, stems, and fiber. Herbaceous plants, which constituted 8%of the plant foods consumed, produced small amounts of fruit (four of five determined species), but may have been more important for the nutrition provided by their young shoots and stems. Fifty-four percent of all identified foods were collected in the central region, while 46%originated in the northeast. Table I lists all plant species, parts consumed, location of collection, habitat, and vegetation types. Fruiting trees are grouped by size; 40% are taller than 20 m, 53% are between 8 and 20 m, and 7% are less than 8 m tall. Since seeds and plant parts were the dietary items most frequently found in the fecal samples, it was not always possible to tell what parts of fruit other than seeds were ingested. The difficult field conditions precluded close observation of actual feeding behavior. Most plant species collected during direct observation were remnants of partially ingested items. The kernel of the nut-like fruit, C o d a edulis (Olacaceae), was consumed, and the hard covering was discarded. In the Myristacaceae, whose aromatic 3-4-cm fruits contain a brightly colored aril enclosing a single seed, the evidence implies that only the seed andor aril was eaten. Small and medium-sized fruits with abundant small seeds dispersed throughout the fruit were usually consumed whole. It appears that ripe fruits are mostly preferred, although unripe fruit appearing a t the beginning of the fruiting season may also be eaten. The unripe pulp of Irvingia gabonensis (wild mango) was partially consumed, and the pit was opened to obtain the single seed. The pulp and seeds of Annonidium mannii, a large, soft fruit 20-30 cm long and often in excess of 10 kg, may provide food for several individuals. Most of the bark, leaves, and stems in the fecal and stomach contents were in minute quantities and sizes or partially digested, and were, therefore, too difficult to distinguish. Blades of the grass Atractocarpa olyraeformis, found in two fecal samples, were large (3-5 cm) and intact, indicating that they were probably not easily digestible. Domesticated Plant Species Native inhabitants reported incidents of mandrills consuming plantation crops, such as Elaeis guineensis (palm nut) and Manihot utilissima (manioc), both of which have been cited as mandrill foods [Sabater Pi, 1972; Jouventin, 19751. Other cultivated plants mentioned were bananas, mangos, pineapples, papayas, and avocados. The major dry season (July through August), when plant foods in the forest are least available, was cited as the time of most frequent mandrill incursions into cultivated areas. Burseraceae Apocynaceae Annonaceae Anacardiaceae Family Pseudospondias longifolia (Engl.) Trichoscypha acuminata Annonidium mannii Engl. & Diels Hexalobus crispiflorus A. Rich. Pachypodanthium barteri (Benth) Hutch Polyalthia suaveolens Engl. & Diels Xylopia staudtii Engl. & Diels Indeterminate species (1) Dictyophleba stipulosa (S. Moore ex Wernham) Pichon Pterotaberna inconspicua (Stapf) Stapf Dacryodes buttneri Engl. Dacryodes edulis (G.Don) Lam Dacryodes klaineana (Pierre) Lam Species 3 - I I I I11 I I11 - I I 1,II I I C N C C N C N N C N N N Fruit pulp Fruit, seeds Entire fruit Ripe, unripe Fruit pulp Fruit, seed Seeds Fruit, seed Fruit, seed Fruit pulp Fruit pulp Fruit pulp Fruit pulp (continued) 3 3 3 1 L 2 2 3 2 2 2 I N Fruit pulp Vegetation type Habitat Location Parts eaten TABLE 1. Identification of Plant Species Consumed by Mandrills: Parts Consumed, Geographical Location, Habitat, and Vegetation Type* Hippocrateaceae Guttiferae Graminae Euphorbiaceae Dilleniaceae Ebenaceae Dichapetalaceae Caesalpiniaceae Family Detarium macrocarpum Harms Dialium pachyphyllum Harms Dialium sp. Guibourtia tessmanii (Harms) J. Leonard Dichapetalum sp. Dichapetalum sp. Tetracera sp. Diospyros cra-ssiflora Hiern Diospyros zenkeri (Gurke) F. White Alchornea firibunda Mull. Arg. Discoglypremna caloneura Prain Uapaca heudelotii Baill. Uapaca sp. Atractocarpa olyraeformis Franch. Pentadesma butyracea Sabine Salacia sp. Species 2 2 3 L L L 2 2 1 3 I - - I I I N C C N N C N N Fruit, seeds Fruit, seeds Fruit pulp Fruit, seed Fruit, seed Seed Fruit pulp Fruit pulp HP 2 L I I - Fruit, seeds Pulp, seeds 2 2 N C - I11 N Fruit, seeds Seeds Grass blade N N 1,II C N Entire fruit Pith of trunk I 3 I C Fruit pulp Vegetation type Location Parts eaten Habitat TABLE I. Identification of Plant Species Consumed by Mandrills: Parts Consumed, Geographical Location, Habitat, and Vegetation Type* (Continued) Myristicaceae Moraceae Mimosaceae Marantaceae Icacinaceae Irvingiaceae Indeterminate species (1) Irvingia gabonensis (Aubry-Lecomte ex ORorke) Baill. Klainedoxa gabonensis Pierre Ataenidia conferta (Benth.) Milne-Redh. Sarcophrynium brachystachy (Benth.) K. Schum Pentaclethra macrophylla Benth. Ficus sp. Ficus sp. Musanga cecropiodides R. Br. Myrianthus arboreus (P. Beauv.) Pycnanthus angolensis (Welw.) Warb. Staudtia gabonensis Warb. N N C C C C Fruit, seeds Fruit, seeds Fruit pods Pulp, seeds Aril, seed Aril, seed? C I 1,II 1,II - (continued) 3 3 2 2 2 HP I11 N 1,II 3 HP I I rq C 3 I Fruit, seeds Fruit, base of young shoots Fruit, base of young shoots Seeds - - C C Seeds Fruit, seeds Aframomum sp. Lophira alata Coula edulis Baill. Diogoa zenkeri (Engl.) Exell & Mend. Heisteria sp. Olax sp. Ongokea gore Pierre Ancistrophyllum sp. Calamus deeratus Mann & Wendl. Deidamia sp. Nauclea didderrichii (DeWild & Th. Dur.) Merrill Blighia welwitschii (Hiern) Radlk. Baillonella toxisperma Pierre Gambeya boukokoensis Gambeya lacourtiana Gambeya sp. Indeterminate species (4) Cissus sp. Aframomum giganteum Species C L HP HP - I1 - - - 3 - C C C N,C C N Seeds Pulp, seeds Seeds Seeds Seeds Fruit, seeds, hearts of stems Fruit, seeds I 3 I C Entire fruit I 3 I N Pulp, seeds L 2 I 2 2 I C C Pulp, seeds Fruit pulp - 3 3 3 Vegetation type 3 2 2 C N N N N Entire fruit Pulp, seed Pulp Entire fruit Entire fruit I I 1,111 Habitat I11 I11 C C C Location Fruit, seed Pulp Pulp, seed Parts eaten *Location: N, North Gabon; C, Central Gabon. Habitat: I, Primary forest; 11, Secondary and Forest Edge; 111, Riparian and Inundated Forest. Tree sizes: 1,less than 8 m; 2, 8-20 m; 3, more than 20 m. L, liana; HP, herbaceous plant; -, undetermined. Vitaceae Zingiberaceae Sapotaceae Sapindaceae Passifloraceae Rubiaceae Palmaceae Ochnaceae Olacaceae Family TABLE I. Identification of Plant Species Consumed by Mandrills: Parts Consumed, Geographical Location, Habitat, and Vegetation Type* (Continued) Diet and Habitat Preference of Mandrill / 17 Vertebrate Foods Vertebrate foods appear to be taken infrequently on an opportunistic basis and probably account for only a small fraction of the total food intake of mandrills. Table I1 lists the number of determined vertebrate foods. All were discovered in the fecal samples. In addition, the remains of a species of terrestrial crab of the family Potamonidae was found in association with mandrill tracks in a streambed, indicating manipulation and perhaps consumption of the crab by one or more mandrills. Nest raiding accompanied by the consumption of birds’ eggs and nestlings has also been reported for mandrills [A. Brosset, personal communication], The low frequencies of animal foods (Fig. 4) suggest that their consumption may have been fortuitous (i.e., occurring in the act of exploratory foraging). Invertebrate Prey The diverse array of invertebrate species consumed further illustrates the mandrill’s eclectic feeding behavior (Table 111). Beetles (Scarabeidae), ants (Formicidae), and spiders (Araneida) were the most frequently consumed items. The eight species of beetles, all dung feeders, construct their galleries directly below the dung pad, which is used as a food source for the developing larvae. Since these were the only species represented in the material examined, it can be safely assumed that the adult beetles were collected on or under the dung rather than a t flowers, and collection thus required inspection and manipulation of dung pads [D. Faulkner, personal communication]. Among the 14 species of ants consumed, the five Odontomachus species and two Polyrachis species were the most numerous. These species live in small colonies and generally nest in dead trees or in soil underneath logs. The Odontomachus species, in particular, are predaceous and aggressive. Other smaller scavenging species such as Solenopsis may have been ingested along with fruit or plants [B. Darchen, personal communication]. All specimens of Araneida prey are large, ground-dwelling, hunting spiders that hide under rocks or leaf litter during the day and can also be found on tree trunks. The remaining invertebrate prey were present in small numbers suggestive of opportunistic feeding. The presence of larva and pupa of three invertebrate species suggests that these nutritionally rich food sources were actively sought, since they are typically found in rotting logs or under soil. Most of these soft-bodied immature stages may have been destroyed during mastication. Diversity of Food Species and Dietary Flexibility Figure 2 illustrates the percentage of occurrence of fruit species and plant foods in mandrill fecal samples and stomach contents. Plant foods included leaves and fibrous plant matter from a variety of sources (stems, bark, petioles, etc.). Six percent of the fecal specimens contained no seeds or evidence of fruiting species, but remnants of plant foods were noted. The greatest percentages of occurrence were in the three to four fruit species category. Plant foods were present in 92% of the stomach contents and in 51% of the fecal material. The percentages of all foods procured from stomach contents may reflect seasonal availability, since 58% of the stomachs were collected from mid-June to midAugust, the major dry season and the critical period for abundance of all foods. Consequently, stomach contents collected during more productive seasons may yield a n increased number of fruit species. Most fecal samples were collected from September to May, when fruit was generally available. The majority of identified fruit species (88%)were small to medium-sized (1-8 cm), with only two species exceeding 10 cm in diameter (Fig. 3). Extremely small 18 / Lahm TABLE 11. Vertebrate Foods in Mandrill Fecal Samples (N = 63)and Stomach Contents (N = 12) Phylum: Chordata Class: Reptilia Order: Chelonia Family: Testudinidae Undetermined species (1)(tortoise) Class: Amphibia Order: Salientia Undetermined species (1)(frog, aquatic?) Class: Mammalia Order: Rodentia Family: Hystricidae Genus: Atherurus africanus (Gray, 1842) (African brush-tailed porcupine) Family: Muridae Genus: Malacomys longipes Milne-Edwards, 1877 (Swamp rat) Order: Insectivora Family: Soricidae Genus: Paracrocidura schoutedeni de Balsac, 1956 (Shrew) L 2 40 30 20 10 0 0 1 2 3-4 5- 7 Plant Foods Number o f Fruit Species Fig. 2. Percentages of occurrence of fruit species and plant foods in mandrill fecal samples (N = 63) and stomach contents (N = 12). Diet and Habitat Preference of Mandrill / 19 50 - 45. 40 - 35 - Fruit Sizes (cm) Fig. 3. Percentages of fruit sizes consumed by mandrills, expressed as diametric measurements in centimeters. loot 9 8 7 - 5 6 :. 5 ul C 3 P' Y 3 2 1 Plants Fruits Insects LL, Fungi Earth Animal Matter Food Categories Fig. 4. Frequencies of occurrence of food categories in mandrill stomach contents and fecal samples. Animal matter includes all noninsect animal foods. 20 I Lahm TABLE 111. Invertebrate Species Consumed by Mandrills Phylum: Mollusca Undetermined species (mollusc) (1) Order: Coleoptera Family: Scarabeidae Tribe: Onthophagini Genus: Onthophagus sp. Diastellopalpus sp. Proagoderus gibbiramus Tribe: Onitini Genus: Onitis sphinx? Onitis subcrenatus? Tribe: Oniticellini Genus: Liatongus interruptus Tribe: Coprini Genus: Copris sp. Catharsius sp. Family: Brentidae Undetermined species (1) Undetermined Coleoptera species (adult) (1) Undetermined Coleoptera species (larva) (1) Order: Hymenoptera Superfamily: Chalcoidea Undetermined species (parasitic wasp) (1) Family: Braconidae Undetermined species (wasp) (1) Family: Formicidae Subfamily: Myrmicinae Genus: Crematogaster clariuentris Crematogaster striatula Crematogaster sp. Solenopsis sp. Myrmicaria sp. Subfamily: Dorylinae Genus: Anoma nigricans Subfamily; Ponerinae Genus: Oclontomachus spp. (5) Subfamily: Formicinae Genus: Polyrachis spp. ( 2 ) Order: Hymenoptera Family: Formicidae Undetermined species (1) (pupa) (continued) fruits such as Discoglypremna caloneura (1cm) were often swallowed whole. Small seeds are probably less susceptible to damage from mastication. Few fragmented seeds were present in the analyzed materials. Invertebrates may be highly valued food items (Fig. 4) as reflected by the presence of insect remains in 100% of the stomach contents and 91% of the fecal samples. It was not possible to determine whether the ingestion of earth (25% frequency) was intentional (geophagy) or incidental. Captive mandrills have been observed eating small particles of soil (author’s personal observation). Relative Intake of Foods Table IV shows the relative importance of the five categories of food found in the stomach contents. The plant category includes leaves, stems, fibrous matter, Diet and Habitat Preference of Mandrill / 21 TABLE 111. Invertebrate Species Consumed by Mandrills (Continued) Subfamily: Dolichoderinae Undetermined species (1) Class: Diplopoda Undetermined species (1)(millipede) Chilopoda Undetermined species (1)(centipede) Arachnida Order: Araneida Family: Sparassidae Undetermined species (1) Theraphosidae Undetermined species (1) Selenopidae Undetermined species (1) Clubionidae Undetermined species (1) Ctenidae Undetermined species (1) Order: Ricinuleida Undetermined species (1) Order: Acari Family: Ixodidae Undetermined species (1)(tick) Order: Lepidoptera Undetermined species (2) (adult, larva) Order: Collembola Undetermined species (1)(springtail) Order: Hemiptera Family: Nabidae Undetermined species (1)(damsel bug?) Reduviidae Undetermined species (1)(assassin bug) Undetermined Hemiptera species (3) Order: Orthoptera Undetermined species (2)(grasshopper) Superfamily: Blattoidea Undetermined species (2) (cockroach) bark, and herbaceous plants. Fruits consist of fruit and seeds. Fruits, plants, and insects formed the most important parts of the diet. Fruit was predominant, accounting for a n average of 88%of dry weight of stomach contents. Plants and insects each averaged 5% of dry weight in the same stomachs. The quantity of fungi was relatively low, but individual values varied, ranging from 1% to 4%. Earth-eating was apparently relatively rare, since earth was found in only three of 12 stomachs, where it averaged 2% of dry weight. Seasonal variability could not be ascertained because most (58%)stomachs were obtained during the 3 months of the major dry season. However, there was a general tendency toward increased consumption of plants and decreased ingestion of fruits and insects from June to August, when the latter are less abundant. Insects made up slightly higher proportions than did plants in most individuals. 22 I Lahm TABLE IV. Relative Intake of Different Food Categories for Mandrillus sphinx Expressed as Percentages of Dry Weight in Stomach Contents (N = 12)" Sex 1.F 2. F 3. M 4. F 5. M 6. M 7. M 8. M 9. F 10. M 11. F 12. M Date Dry wt. (g) Plants Fruits Insects Fungi Earth Dec 10 Dec 10 7.1 7.4 59.2 15.7 100.2 116.5 259.3 446.0 81.1 178.4 62.0 140.0 .04 .03 .04 .06 .09 .07 .10 .I1 .03 .02 .01 .03 .05 .89 .90 .91 .84 .86 .81 .05 .04 .05 .09 .05 - .02 .03 - - May 18 Jun 21 Jul15 Jul23 Aug 04 Aug 11 Aug 16 Aug 16 Aug 23 Sep 09 Averages .82 .84 .91 .90 .97 .92 .88 .08 .08 .05 .04 .04 .02 .05 .05 - .01 - .04 - .01 .03 .01 - - .02 .02 - *M, male; F, female. Diet and Habitat It was possible to determine the habitat of 77%of the 63 identified plant species. The remaining 23% were difficult to identify beyond the familial or generic taxonomic levels. Of the 49 species determined a t the specific level, 71%occur in primary forest, 14%in secondary and forest edge, and 14%in riparian and inundated forests. Considering only fruiting tree species, i.e., excluding the 11 species of lianas and herbaceous plants that also provide fruits, 74% occur in primary forest, 14% in secondary and forest edge, and 12% in riparian and inundated forest habitat. These percentages strongly suggest that mandrills in this region forage primarily in areas associated with mature forest tree species but also take food in other habitats. No plants were typical of savannahs, which exist in the central study region. This evidence is corroborated by the predominance of mandrill sightings in this region on hillsides and plateaus in primary and secondary forests relatively free of dense undergrowth and close to streams. Much of the secondary forest a t lower elevations had a dense undergrowth of herbaceous plants from ground level to 3 m. Mandrills were never observed in this habitat, nor on the savannah, although signs (tracks, dung) of their crossing narrow expanses of savannah were noted. The condition of the evidence suggested that the crossings occurred in the early morning and late afternoon, when sunlight intensity was minimal. In the northeastern study area, mandrills were seen only on plateaus and slopes above streams or rivers where the undergrowth was less dense. Mandrill tracks were also found along and in stream beds, indicating that groups forage in riparian habitat in addition to obtaining water there. Sixty percent of identified animal foods came from this habitat (Table 11). DISCUSSION Dietary Diversity and Forest Regeneration The importance of primates as seed dispersers has been widely documented [Hladik & Hladik, 1967; Liebermann et al, 19791. The mandrill's selection of a diverse array of mature, fleshy fruits of which both seeds and pulp are consumed suggests that it is a n important seed disperser for many plant species. CharlesDominique  has demonstrated a complex dispersion of tree species in the Diet and Habitat Preference of Mandrill I 23 tropical rainforest of Gabon. Fruiting trees are extremely scattered but yield an abundant though localized crop. Those with a meager production of fruit are abundant, while those producing extensive amounts of fruit are rare and widely scattered. If mandrill groups adapt their social organization and foraging strategies to the seasonal availability of foods, their selection of a wide range of foods within a large supplying area would disperse a variety of plant seeds over considerable distances. Since fruits constitute the major portion of the mandrill’s diet, the large group size and home range of this species suggest that mandrills may be important seed dispersal agents. Hoshino et a1  estimated home-range sizes of mandrills in Cameroon as 5 km2 for a group of 15 and 28.5 km2 for a group of 95. In contrast, the home-range sizes of six sympatric arboreal primate species in Gabon vary from .1 to 2.0 km2 [Gautier-Hion, 19781. Significance of Food Selection The dietary choices and distribution of primate species may be affected by high concentrations of toxins in plants [McKey, 19781. As Hladik [1977,1981]has shown, plant species eaten by primates in Gabon vary widely in digestibility. My observations suggest that mandrills may select items of low toxicity, and that their foraging strategies and diet may contrast in this respect with the feeding habits of the chimpanzee. Both mandrills and chimpanzees are large, frugivorous, semiterrestrial foragers with large home ranges. Hladik  reported a 28% annual proportional intake of leaves and stems by chimpanzees in terms of fresh weight, as compared to 5% averaged for mandrill stomach contents in the present study. The mandrill’s digestive system may not be equipped to efficiently process coarse vegetation such as leaves, bark, and stems. These foods were present in 51% of the mandrill fecal samples, which suggests that many plants are not fully digestible (Fig. 3). The great diversity of invertebrate species consumed by the mandrills (Table III) may reflect the mandrill’s need for specific dietary components not obtainable from vegetable foods. Although no information is available about mandrill feeding time budgets, the data presented here indicate that invertebrates of many species (including 14 species of ants) are regularly eaten. As Hladik  has demonstrated, such insects are rich sources of amino acids. Chimpanzees in Gabon spent 30-50% of feeding time foraging for insects. Hamilton et a1  observed that animal matter is the preferred food of Papio ursinus, especially when insects are seasonally abundant. The seasonal fluctuation in availability of foods is probably an important determinant of the diet of mandrills. Redford and Dorea  demonstrated that some of the most nutritionally valuable insect forms (larvae, alate ants, and termites) are only seasonally available. Although fruit and insects may require a greater expenditure of time and energy to obtain, the ubiquitous leaves and plant parts may be less preferred foods owing t o the presence of digestion-inhibiting compounds in green plants. Vertebrate prey also may be an important, though relatively infrequent, source of scarce nutrients for mandrills, as has been also suggested for baboons [Hausfater, 1976; McGrew et al, 1978; Strum, 19801 and chimpanzees [Morris & Goodall, 19771. Feeding Behavior and Social Organization Habitat may be selected for availability of preferred foods and sleeping sites, all of which, along with other environmental variables and phylogenetic characters, probably affect the mandrill’s social structure and population dynamics [Lahm, 19851. Although several habitat types were surveyed, the fact that most data collection and observations occurred in forested areas free of very dense undergrowth 24 I Lahm suggests that mandrills prefer this type of environment, possibly because travel, visibility of, and contact with group members, and detection of potential predators are facilitated. Both home-range and group sizes of mandrills may be dependent on resource stability. The patchy distribution of fruiting trees in the rainforests of Gabon may result in the common usage of a large home range by several mandrill groups, which appear to form larger aggregations on a seasonal basis [Lahm, 19851. If a number of mandrill groups occupy one large or several smaller overlapping home ranges, groups may employ different travel patterns to resource sites, thus reducing intergroup competition, as have been suggested for hamadryas baboons [Sigg & Stolba, 19811. Intratroop competition for abundant, but dispersed foods may be affected by age or sex-specific dietary preferences [Gautier-Hion, 19801. Although adult male and female mandrills were observed foraging in trees, smaller group members more frequently engaged in this activity and were also more likely to climb trees in flight. Rowel1 119661 observed that immature olive baboons had greater access to nutritionally rich immature leaves and shoots owing to their greater agility in trees. If juvenile and subadult mandrills are more easily able to utilize tree leaf materials, adults may benefit by eating comparatively less toxic plants in the herb layer of the forest [Waterman, 19841. Interspecific Competition Mandrills share their rainforest habitat with a myriad of species, including 120 species of mammals and over 200 species of birds. Competition for food and ecological niches is mitigated by the evolution of nocturnal and diurnal activity patterns. Seventy percent of the mammalian species are nocturnal, and 96% of the birds are diurnal [Charles-Dominique, 19751. Excluding the avian fauna, some of which also consume fruit, 65 species of mammals, whose diets chiefly comprise fruit, leaves, stems, flowers, and bark [Emmons et al, 19831, compete with the mandrill for fruit, its major dietary item. Differential spatial and temporal utilization of the habitat may decrease competition for resources [Gautier-Hion et al, 19811. The majority of mandrill foods identified in the present study are found in primary forest, while secondary forest, riparian, and inundated habitats provide important supplementary foods. Analysis of the diets of the primates, ruminants, and sciurids of Gabon [Gautier-Hion et al, 19801 revealed that fruit, leaves, and animal matter collectively accounted for 98% of all ingested food. While over half of the identified fruit species exploited by one taxonomic group were also exploited by one or more of the other taxa, each taxon’s general preference for different fruit parts mitigated excessive competition. In general, primates prefer fruit pulp, ruminants swallow small whole fruits, and sciurids eat seeds. Of the primates in Gabon, only mandrills, gorillas (Gorilla g. gorilla), and chimpanzees (Pan t. troglodytes) engage in extensive terrestrial foraging [Emmons et al, 19831. Although fruits are important food sources for these three species, the apes are more folivorous than mandrills [Gautier-Hion et al, 19801. Interspecific competition also may be reduced through choice of alternative foods during times of environmental periodicity [Hamilton et al, 19781. The mandrill’s relatively large body size, broad, elongated hands and feet, and enlarged incisors may offer advantages related to a n increased capability for processing and exploiting a variety of acceptable foods in both arboreal and terrestrial ecological niches. This species’ diverse selection of foods within a large supplying area permits the efficient utilization of a complex environment. Diet and Habitat Preference of Mandrill / 25 CONCLUSIONS 1. The mandrill diet consists primarily of fruit, supplemented by various other plant parts, numerous invertebrate species, and occasional vertebrates. 2. The high frequency of insect consumption by mandrills suggests that invertebrates may be highly valued food items. 3. Although mandrills are basically terrestrial, they probably spend a greater percentage of time foraging in trees than do the Papio species. 4.Mandrills appear to prefer primary forest, but they also forage in secondary, riparian, and inundated forests. 5. Since mandrills select numerous fruit species within large supplying areas, they may play a significant role in seed dispersal. 6. Competition for resources in the rainforest is mitigated by species’ preferences for particular food parts and by differences in spatial and temporal habitat utilization. ACKNOWLEDGMENTS I especially thank W.C. McGrew for providing the opportunity for me to pursue the mandrill field study, and Messrs. Paul Posso and Raphael Dipouma for permission to work in Gabon. The following individuals are gratefully acknowledged for identification of dietary items: Y. Cambefort, B. Darchen, G. Dubost, D. Faulkner, F. Feer, A. Hladik, R. Letouzey, and A. Moungazi. Thanks are due to Drs. Clifford Jolly, Donald Lindburg, and Oliver Ryder for much appreciated constructive criticism of the manuscript, to Sharon Dinwiddie for preparing it for publication, and to Stephen Nash for permission to use his map of Gabon. I also extend my heartfelt appreciation to innumerable people in Gabon, France, and San Diego for their assistance, moral support, warmth, and hospitality during the past 3 years. REFERENCES Caballe, G. Essai sur la Geographie Forestiere du Gabon. ADANSONIA, Series 2, 17~425-440,1978. Charles-Dominique, P. Nocturnality and diurnality: An ecological interpretation of these two modes of life by a n analysis of the higher vertebrate fauna in tropical forest ecosystems, pp 69-90 in PHYLOGENY OF THE PRIMATES. P. Luckett; F. Szalay, eds. New York, Plenum Press, 1975. de Saint Aubin, G. La foret du Gabon. PUBLICATION NO. 21, Centre Technique Forestier Tropical, Nogent-sur-Marne, 1963. Dorst, J.; Dandelot, P. A FIELD GUIDE TO THE LARGER MAMMALS OF AFRICA. London, Collins, 1970. Emmons, L.H.; Gautier-Hion, A.; Dubost, G. Community structure of the frugivorous-folivorous mammals of Gabon. JOURNAL OF ZOOLOGY, LONDON 199:209-222,1983. Florence, J.; Hladik, A. Catalogue des phanerogames et des pteridophytes du nord-est du Gabon (sixieme liste). ADANSONIA, Series 2, 20:235-253, 1980, Gautier-Hion, A. Food niches and co-existence in sympatric primates in Gabon, pp 269-286 in RECENT ADVANCES IN PRIMATOLOGY, VOL. I BEHAVIOR. D.J. Chivers; J. Herbert, eds. London, Academic Press, 1978. Gautier-Hion, A. Seasonal variations of diet related t o species and sex in a community of cercopithecus monkeys. JOURNAL OF ANIMAL ECOLOGY 49~237-269,1980. Gautier-Hion, A.; Emmons, L.H.; Dubost, G. A comparison of the diets of three major groups of primary consumers of Gabon (primates, squirrels and ruminants). OECOLOGIA (Berlin) 45182-189,1980. Gautier-Hion, A.; Gautier, J.P.;Quris, R. Forest structure and fruit availability as complementary factors in influencing habitat use by a troop of monkeys (Cercopithecus cephus). LA TERRE ET LA VIE 35x511536,1981. Grubb, P. Distribution, divergence and speciation of the drill and mandrill. FOLIA PRIMATOLOGICA 20:161-177,1973. Hamilton, W.J. 111; Buskirk, R.E.; Buskirk, W.H. Omnivory and utilization of food resources by Chacma baboons, Papio ursinus. AMERICAN NATURALIST 112:911-924, 1978. Hausfater, G. Predatory behavior of yellow baboons. BEHAVIOR 56:440-468,1976. Hladik, A. Phenology of leaf production in rainforest of Gabon: Distribution and composition of food for folivores, pp 51-73 in THE ECOLOGY OF ARBOREAL FOLIVORES. G.G. Montgomery, ed. Washing- 26 1 Lahm ton, D.C., Smithsonian Institution Press, 1978. Hladik, C.M. Chimpanzees of Gabon and chimpanzees of Gombe: Some comparative data on the diet, PP 481-503 in PRIMATE ECOLOGY. T.H. Clutton-Brock, ed. London, Academic Press, 1977. Hladik, C.M. Feeding strategies of forest primates, PP 215-253 in OMNIVOROUS PRIMATES, GATHERING AND HUNTING IN HUMAN EVOLUTION. R.S.O. Harding; G. Teleki, eds. New York, Columbia University Press, 1981. Hladik, C.M.; Hladik, A. Observations sur le r81e des primates dans la dissemination des vegetaux de la for& Gabonaise. BIOLOGICA GABONICA 3~45-58,1967. Hoshino, J.; Mori, A.; Kudo, H.; Kawai, M. Preliminary report on the grouping of mandrills (Mandrillus sphinx) in Cameroon. PRIMATES 251295-307,1984. Jolly, C. The large African monkeys as an adaptive array, pp 139-175 in OLD WORLD MONKEYS, EVOLUTION, SYSTEMATICS AND BEHAVIOR. J.R. Napier; P.H. Napier, eds. New York, Academic Press, 1970. Jouventin, P. Observations sur la socio-ecologie du mandrill. EXTRAIT TERRE ET LA VIE. REVUE D’ECOLOGIE APPLIQUEE 29:493-532,1975. Lahm, S.A. The Ecology of the Mandrill, Mandrillus sphinx, Ritgen 1824. Unpublished master’s thesis, San Diego State University, 1985a. Lahm, S.A. Mandrill ecology and the status of Gabon’s rainforests. PRIMATE CONSERVATION 6:32-33,1985b3. Lieberman, D.; Hall, J.B.; Swanke, M.D.; Lieberman, M. Seed dispersal by baboons in the Shai Hills, Ghana. ECOLOGY 60:6575, 1979. McGrew, W.C.; Tutin, C.E.G.; Baldwin, P.J.; Sharman, M.J.; Whiten, A. Primates preying upon vertebrates: New records from West Africa. CARNIVORES 1:42-45, 1978. McKey, D. Soils, vegetation and seed-eating by black colobus monkeys, pp 423-437 in THE ECOLOGY OF ARBOREAL FOLIVORES. G.G. Montgomery, ed. Washington, D.C., Smithsonian Institution Press, 1978. Moreno-Black, G. The use of scat samples in primate diet analysis. PRIMATES 19:215221, 1978. Morris, K.; Goodall, J. Competition for meat between chimpanzees and baboons of the Gombe National Park. FOLIA PRIMATOLOGICA 28:109-121,1977. Napier, J.R.; Napier, P.H. A HANDBOOK OF LIVING PRIMATES. London, Academic Press, 1967. Redford, K.H.; Dorea, J.G. The nutritional value of invertebrates with emphasis on ants and termites as food for mammals. JOURNAL OF ZOOLOGY, LONDON 203:385-395,1984. Rowell, T. Forest living baboons in Uganda. JOURNAL OF ZOOLOGY, LONDON 149:344-364,1966. Sabater Pi, J. Contribution to the ecology of Mandrillus sphinx, Linnaeus 1758 of Rio Muni (Republic of Equatorial Guinea), FOLIA PRIMATOLOGICA 17:304-319, 1972. Sigg, H.; Stolba, A. Home range and daily march in a hamadryas baboon troop. FOLIA PRIMATOLOGICA 36:40-75.1981. Strum, S.Processes and products of change: Baboon predatory behavior at Gilgil, Kenya, pp 193-214 in OMNIVOROUS PRIMATES, GATHERING AND HUNTING IN HUMAN EVOLUTION. R.S.O. Harding; G. Teleki, eds. New York, Columbia University Press, 1980. Waterman, P.G. Food acquisition and processing as a function of plant chemistry, pp 177-211 in FOOD ACQUISITION AND PROCESSING IN PRIMATES. D. Chivers; B. Wood; A. Bilsborough, eds. New York, Plenum Press, 1984. Wolfheim, J. PRIMATES OF THE WORLD: DISTRIBUTION, ABUNDANCE AND CONSERVATION. Seattle, University of Washington Press, 1984.