Peter S. Rodman
Department of Anthropology
University of California, Davis
Davis, California 95616

The rationale for a comparative approach to reconstructing the earliest diets of humans is clear. Orangutans, chimpanzees, bonobos and gorillas are the closest living relatives of humans, and by the usual logic of phylogenetic inference, some sense should be made of diet of the common ancestor with reference to the shared characteristics of diets of those relatives. A great deal has been written about feeding of the apes, and virtually every field study has produced at least one list of the plant species and parts taken by the subjects. Although there are other important comparisons to be made of the diets of the apes, this paper addresses the taxonomic overlap of the known plant diets as a step towards understanding any common features of those diets. As Tutin (1994), who has made an incomparable contribution to our knowledge of plant food species of the African apes, wrote,

The first step in comparison is to sort out the taxa taken by the apes from those that are not, and those parts that are taken from those that are not. This presentation is a simple beginning to sorting the data, primarily at the family level because there is considerable information on characteristics of plant families to use for a first, or even second, cut at discovering details of the diets of the apes and other primates.

If there is any original contribution to this work, it is the reconciliation and standardization of the classifications of plants provided in the literature on diets of the apes before evaluating characteristics of the families of food plants. Several important analyses of the qualities of plant foods of the apes have been completed, and the animal component of the diets of the great apes is quite important, if only because of the significance of meat to some--but not all--of us. There is a discouraging word at the end regarding phenotypic flexibility in diets that probably precludes useful inference about ancestral dietary adaptations.

A list of all plants identified in 16 data sets for 15 sites was assembled from primary sources (Table 1).

Table 1. Sources of data and studies of the great apes covered in this analysis

The plant parts recorded were coded in eight categories as eaten (1) or not (0) according to the reports and records of the studies cited. These are:

• Fruits, variously labeled as pulp or fruit
• Seed
• Flowers (=blossoms)
• Leaf (= mature, immature, petiole)
• Pith (including "stem" for some lists)
• Bark (including "cambium" and "twigs" in some lists)
• Other (wood, tubers, galls)
• Other 2 (including any record of seed, flower, pith or other plant component; this category was used because some field workers have been less able or willing to differentiate between the seed and the fruit)

The individual records for plant species each contain columns for all the apes (in the primary records, columns are present for each plant part for each species of plant, for each of the apes, even when only one ape species has taken that plant/part) for purposes of summation. The summations over taxa fill in columns for the different apes as the whole data set is reduced first to all represented species (from the whole set, which includes records for a species for each site at which it was reported), then to all represented genera, and finally to all represented families.

The taxonomic data set was cleaned as follows: First the complete list was sorted by genus, species, and family as identified in the original data sets. Using various approaches, specific and generic names given were reconciled between studies and corrected by eliminating obvious spelling errors in the original sets. In some cases, such errors were clear from, for example, the presence of a particular species in multiple African sets in which the name from one study differed from that in others by a single letter. The accepted spellings of names of families, genera and species were checked as carefully as possible against information in several standard databases of plant systematics:

The availability of all these materials on the WWW allowed efficient cross checking among databases, including the one produced here. In the case of synonymies, the generic name indicated as the current accepted name in the ING is used, and all spellings of generic names are reconciled with that list.

Table 2: Numbers of taxa identified as food sources in each study or site*

There is a reasonably large set of families shared between the two major floral regions (Fig. 1). This is not particularly surprising since most of the families are primarily pan tropical in distribution (Watson and Dallwitz, 1992-)

Table 3: Proportions of all identified taxa from which plant parts are taken

These data can be examined most conveniently at the family level as in the following comparison of overlap in families of the plant sources of foods of the four apes.

There is highest overlap between gorilla and chimpanzee in the whole set of families identified. The high overlap between the latter species (note in green, Fig. 2) is partially accounted for by the fact that three data sets for each are for areas where gorillas and chimpanzees are sympatric (Lope, Ndoke and Kahuzi). Removing the data from Ndoke, Lope, and Kahuzi changes the result a bit, so that the pairwise overlap in families of plants is more uniform (blue in Fig. 2)

I have chosen the family level though there might be reason to make comparisons at the generic level among the African apes, which share floral reason. The following figure shows that overlap of genera among the three African apes is generally lower than at the family level. The higher overlap of gorilla and chimpanzee is due to the three studies of these species in sympatry.

There are interesting, but not surprising variations and similarities among the great apes in the proportions of plant parts taken from each subset of families.

The figure suggests that gorillas use fruit from a smaller proportion of plants than other apes and leaves from a greater proportion of plant families. To the extent that the manner of reporting the data is comparable, this contrasts, with others' interpretations of the results of studies of lowland gorillas. Since the data set includes plants of mountain gorillas of the Ruwenzori Mountains where there is virtually no fruit, Fig. 1 shows data only from lowland sites as well. The contrast with the other apes remains. The data show that gorillas take leaves (and bark) from a larger proportion of the families of plants used than other great apes. While the proportions of plant taxa from which food is selected do not directly represent dietary compositions, it is interesting that the combined data sets show this distinct difference of gorillas from the other apes. The relationship of proportions of leaf sources relative to fruit sources at this level are the opposite for gorillas from those of the other three great apes.

In order to arrive at a set of families that may be preferred or avoided by the apes as a whole, I have entered the numbers of genera per family in the world obtained from the FFP database. I then regressed the number of genera per plant food family of the apes on the number of genera of those families in the world. This is not a tremendously sound approach, since the distribution of the world genera may be such that the numbers do not reflect availability of genera of the families in Africa and Asia. The (Log-Log) regression is shown in Fig. 5.

The residuals of the regression provide a list of 16 families to be examined further for characteristics that may indicate the reason for their over- and under- representation in the sample of food plants of the apes.

Table 3. Families falling more than one standard deviation from the regression in Fig. 5*

This examination is in progress.

Reconciling and correcting the names of plant species and higher taxa from multiple studies has been tedious, but it is the efforts of the field workers that have been the greatest. I thank all authors of field reports cited here for the years of work that have produced the information summarized here, for their care in collecting specimens, for the suffering it takes to get the samples, and for their persistence at obtaining identifications of the plants. I am also grateful to those who have made data available over the internet, beginning with the set of plant identifications for five African sites provided by Kuroda, Mitani and Tutin and continuing with all the keepers of data sets to which there are links in this presentation. I am indebted to a number of current and former students, including particularly Mark Leighton, Anthony Di Fiore, and Larry Dew, for their persistent and expert work on the botanical side of primatology. Their experience has been instructive for me, and puts my own early work in the shadow of their expert field ecology. Finally, I am delighted to acknowledge Jonathan Greenberg's assistance with the graphics in this presentation.

Badrian, N. and Malenky, R. K. 1984. Feeding ecology of Pan paniscus in the Lomako Forest, Zaire. Pp. 275-299 in The Pygmy Chimpanzee, R. L. Sussman, Ed. Plenum Press, New York.

Galdikas, B. M. F. 1978. Orangutan Adaptation at Tanjung Puting Reserve, Central Borneo. Doctoral Dissertation, University of California, Los Angeles.

Idani, G., Kuroda, S., Kano, T., & Asato, R. (1994). Flora and vegetation of Wamba Forest, Central Zaire with reference to bonobo (Pan paniscus) foods. Tropics 3: 309-332.

Leighton, M. 1993. Modeling dietary selectivity by Bornean orangutans: Evidence for integration of multiple criteria in fruit selection. Int. J. Primatology 14: 257-313.

MacKinnon, J. 1974. The behaviour and ecology of wild orang-utans (Pongo pygmaeus) Animal Behaviour 22: 3-74.

McGrew, W., Baldwin, P.J., and Tutin, C. E. G. 1988. Diet of wild chimpanzees (Pan troglodytes verus) at Mt. Asserik, Senegal: I. Composition. Amer. J. Primatol. 16: 213-226.

Moore, J. (1994) Plants of the Tongwe East Forest Reserve (Ugalla), Tanzania. Tropics 3: 333-340.

Moutsambote, J.-M., Yumoto, T., Mitani, M., Nishihara, T., Suzuki, S., and Kuroda, S. (1994) Vegetation and list of plant species identified in the Nouabale-Ndoki Forest, Congo. Tropics 3: 277-293.

Nishida, T., and Uehara, S. 1983. Natural diet of chimpanzees (Pan troglodytes schweinfurthii): Long term record from the Mahale Mountains, Tanzania. African Study Monographs 3: 109-130.

Remis, M. J. 1994. Feeding Ecology and Positional Behavior of Western Lowland Gorillas (Gorilla gorilla gorilla) in the Central African Republic. Doctoral Dissertation, Yale University, New Haven, CT.

Rijksen, H. D. 1978. A Field Study on Sumatran Orangutans (Pongo pygmaeus abelii Lesson, 1827): Ecology, Behaviour and Conservation. H. Veenman & Zonen, Waganingen.

Rodman, P. S. 1973. Synecology of Bornean Primates. Doctoral Dissertation, Harvard University, Cambridge, MA.

Rodman, P. S. 1988. Diversity and consistency in ecology and behavior. Pp. 31-51 in Orang-Utan Biology, J. H. Schwartz, ed. Oxford University Press, Oxford.

Tutin, C. E. G., White, L. J. T., Williamson, E. A., Fernandez, M., & McPherson, G. (1994) List of Plant Species Identified in the Northern Part of the Lope Reserve, Gabon. Tropics 3: 249-276.

Watts, D. P. 1984. Composition and variability of mountain gorilla diets in the Central Virungas. Amer. J. Primatol. 7: 323-356.

Wrangham, R. W. 1975. The Behavioural Ecology of Chimpanzees in Gombe National Park, Tanzania. Doctoral Dissertation, University of Cambridge.

Yumoto, T., Yamagiwa, J., Mwanza, N., & Maruhashi, T. (1994) List of Plant Species Identified in Kahuzi-Biega National Park, Zaire. Tropics 3: 295-308.