PRIMATES,32(1): 1-7, January 1991

1

Adaptation to Grass-eating in Gelada Baboons R. I. M. DUNBAR and UTPAUL BOSE

University College London

ABSTRACT. Gelada faecal samples were analyzed for nutritional content and for particle size, and compared with similar data for Papio baboons, cattle, and zebra. Particle size in gelada is similar to that for zebra, larger than that for cattle and smaller than that for baboons. Gelada and baboons are less efficient than ungulates at extracting protein from their diet. The data on energy extraction are less easy to interpret and appear to be confounded by dietary and seasonal factors. It is suggested that gelada may be too large to compete effectively with ruminants in low altitude grassland habitats under the climatic conditions that have prevailed in eastern Africa since the late Pleistocene. Key Words: Gelada baboons; Nutritional ecology; Food processing adaptations.

INTRODUCTION Gelada baboons (Theropithecus gelada) are unique among the primates in their exploitation of a graminivorous (grass-eating) niche. Although derived from a stock whose other extant descendents (baboons and macaques) are all frugivores, the theropithecine lineage shows evidence of having evolved specializations for a graminivorous diet at a very early stage in its evolutionary history (JOLLY, 1972; ECK & JABLONSKI,1987; LEE-THORP et al., 1989). Foliage imposes major constraints on the digestive system since a high proportion of the nutrients are locked up in the relatively indigestible cell walls (composed largely of celluloses and lignins: PHILLIPSON, 1977; VAN SOEST, 1982; STEVENS, 1988). Most herbivores have evolved digestive strategies to deal with this problem. The primary options involve either the exploitation of specialized microbial fauna in the forestomach (the ruminant or pseudo-ruminant strategy) or the somewhat less efficient hind-gut fermentation. The former strategy is characteristic of all ruminant mammals as well as those primates that are primarily folivores (e.g. Colobus, Presbytis, Alouatta: BAUCHOP & MARTUCCI, 1968; KAY et al., 1976; MILTON& MCBEE, 1982). The latter strategy is characteristic of the nonruminant ungulates (e.g. horses and their allies) and those species of birds that graze (e.g. geese): it involves an emphasis on bulk feeding in which intake is increased to compensate for a reduced nutrient extraction rate (JANIS, 1976). The problem presented by the gelada is how a non-ruminant derived from a frugivorous ancestral stock manages to cope with a purely folivorous diet. Assuming that the gelada have not yet evolved a symbiotic micro-fauna that undertakes foregut or caecal fermentation on its behalf, the only two options available are (1) to masticate the ingesta so thoroughly that the cells walls are broken down into smaller fragments for easier digestion by conventional means and (2) to ingest a larger quantity of food in order to offset reduced nutrient extraction rates. We know that the gelada do not in fact spend significantly more time feeding than Papio baboons in areas where the two taxa are sympatric (DUNBAR& DUNBAR, 1974). Gelada dentition does, however, differ in significant ways from that of

2

R.I.M.

DUNBAR & U. BOSE

Papio baboons: gelada have relatively smaller incisors and larger molars (chewing teeth) than Papio (JOLLY, 1972; SZALAY& DELSON, 1979) and their molar teeth exhibit distinct adaptations to resist wear and improve their ability to macerate fibrous foods (hypsodonty and deep surface relief) (JOLLY, 1972; EcK & JABLONSKI, 1987). These features suggest that the theropithecines have evolved a dentition adapted to grinding rather than nibbling. In an attempt to explore this question in more detail, faecal samples were examined to determine (1) the efficiency with which gelada are able to comminute vegetation and (2) the extent to which they are able to extract nutrients from their food. To provide a baseline for comparison on both measures, faecal samples for baboons (Papio anubis), zebra (Equus burchelh), and free-ranging cattle were also collected. If gelada are essentially primate horses pursuing a bulk feeding strategy (with or without caecal fermentation), they should be more efficient than baboons but less efficient than cattle, and similar to the zebra.

METHODS Faecal Samples were obtained from free-living populations o f all four species. Those for gelada were collected by Dr. M. J. S. HARRISON in the vicinity of Ankober, Ethiopia (altitude 3,050 m asl) in June 1986 (early wet season). Those for the remaining three species were collected by R.I.M.D. at Gilgil, Kenya (altitude 1,800 m asl) in March 1982 (just after the main rains). Faecal samples were all sun-dried in the field. Nutrient content of the graze may be somewhat higher at Ankober than Gilgil due to higher rainfall and a cooler climate (Table 1). In addition, by the time the Gilgil samples were collected, the grasses had already begun to dessicate, and this may have affected the digestibility of the graze (see BRAUN, 1973). Gelada, zebra, and cattle eat little except grass. Although the Gilgil baboons do eat a significant quantity of grass (HARDING, 1976), Papio baboons are frugivores by preference and never rely wholly on grasses (see, for example, DUNBAR, 1984). Thus, although the baboon samples were collected at a time when the animals were eating grass blades, the baboon data will inevitably be confounded by a higher intake of relatively low-fibre foods (seeds and roots). In order to determine particle sizes, faecal samples were soaked thoroughly to allow fibres to be separated easily and then examined under a dissecting microscope at x 16 power. Three samples were taken from each species. In each case, approximately 100 particles chosen at random were carefully removed from the substrate and their lengths measured against a graticule at a resolution of 17 units to 1 mm. Analyses of the nutritional content of the faeces was carried out at the Nutrition Unit o f the London School of Hygiene and Tropical Medicine. A Kjeltec Auto 1030 Analyzer

Table 1. Crude protein content of grass at Gilgil (Kenya) and at Sankaber Crude protein (%) Gilgil Sankaber Wet season 9.3 -Dry season 8.6 12.5 Altitude (m) 1800 3300 Rainfall (ram/year) 690 1385 Gilgil: R. DUNBAR(unpubl. data); Sankaber and Gich: IWAMOTO(~ DUNBAR(1983).

and Gich (Ethiopia). Gich 10.0 7.0 3900 1465

Grass-eating in Gelada

3

was used to measure nitrogen content; protein was then determined by applying the standard conversion factor of 6.25. Energy content was determined using a ballistic bomb calorimeter calibrated against sucrose (whose thermochemical value was taken to be 16.51 k J/g). For each of the analyses, a quantity of faecal material was homogenized and four samples taken. Some problems were experienced with the calorimeter during the analyses of the baboon and gelada material, which gave rise to some unexpectedly variable results. Extreme values were omitted from the analyses, but, as a check on this, a second set of samples was analyzed some months later.

RESULTS Protein contents of the dung samples are given in Table 2. A Kruskal-Wallis analysis of variance indicates that there are significant differences between the species (H=40.22, p < 0.001). A Scheff6 multiple comparisons test was then used to determine how the species differed from each other. This revealed that both primate species differ significantly from both ungulate species (p < 0.001), but that the gelada did not differ significantly from the baboons nor the zebra from the cattle (p>0.05 in each case). Energy contents for the two series of faecal samples are given in Table 3. There are significant differences between the species on the first series (Kruskal-Wallis analysis of variance: H=76.1, p<0.001), but not for the second (H=10.982, p > 0 . 9 0 ) even though the species' values are of about the same magnitude. A Scheff6 test on the first series indicates that all the species differ significantly from each other (17< 0.01) except cattle and zebra (/7>0.50). Taken together, however, it is clear that although the two baboon species have lower energy contents than the two ungulates, the differences are not in fact that great (at least by comparison with the results for protein content). Comparison of the distribution of fibre lengths for the three samples for each species indicates that there are no within-species differences (gelada: X2=11.58, d f = 8 , p>0.10; baboons: X2=10.62, d f = 8 , p > 0 . 2 0 ; zebra: X2=12.58, d f = 8 , p>0.10; cattle: X2=4.45, d f = 6 , p > 0.50). The samples for each species were therefore pooled and overall distribu-

Table 2. Crude protein content o f faecal samples. Crude protein (o7o) Species

Mean

S.D.

N

Gelada Baboons Zebra Cattle

6.00 6.13 1.75 1.73

0.20 0.41 0.05 0.04

4 4 4 4

Table 3. Energy content of faecal samples. Energy (k J/g) Sample 1

Sample 2

Species

Mean

S.D.

N

Mean

S.D.

N

Gelada Baboons Zebra Cattle

12.71 14.59 16.84 16.76

0.29 0.39 0.57 0.66

4 3 4 4

14.19 13.33 16.42 15.55

0.29 0.27 2.19 0.61

4 4 8 4

4

R . I . M . DUNBAR& U. BOSE

Table 4. Distribution of particle size (maximum length) for each species. Frequency Particle size (mm) Gelada Baboon Zebra < 0.6 56 2 44 0.6- 1.1 104 17 84 1.2- 1.7 73 52 77 1.8-2.3 28 57 41 2.4-2.8 16 35 29 2.9-3.4 14 36 11 3.5 -4.0 6 27 9 4.1-4.6 3 28 3 4.7-5.2 3 18 5 5.3-5.8 2 10 1 5.9+ 2 38 5

Cattle 70 140 63 29 7 6 6 1 1 3 1

Total

307

320

309

327

Mean length (mm) S.D.

1.41 1.12

3.34 1.91

1.61 1.13

1.19 0.94

tions of particle length computed (Table 4). A comparison of the distributions for all four species shows that there are significant differences among them (4 x 6 X2=431.8, df=15, p < 0.001). Partitioning the degrees of freedom indicates that there is no significant difference between the gelada and the zebra (X2= 8.813, d f = 5, p > 0.10), but that these two species when combined differ significantly from both the cattle (X2= 32.126, d f = 5, p > 0.001) and the baboons (X2=260.007, df--5, p < < 0.001). Thus, the degree to which gelada and zebra are able to comminute their food is very much better than Papio baboons, but not quite as good as cattle.

DISCUSSION The results suggest that the gelada are able to break down their food into finer fragments than baboons do, but less competently than cattle who are able to reduce particle sizes significantly more by virtue of being able to chew over their ingesta a second time during rumination. This was as predicted by the fact that the molar teeth of the gelada are more adapted to herbivory than those of conventional baboons (JOLLY, 1972; ECI( & JABLONSKI, 1987). Providing that both nutrient intakes and the digestibilities of the various nutrients are identical, then the nutrient residue in the faeces provides us with a crude measure of the efficiency with which the animals are able to extract nutrients from their food. Faecal crude protein has, in fact, been shown to provide a reliable estimate of dietary crude protein in both bovids and cervids (SINCLAIR, 1977; LESLIE • STARKLY, 1985). In the present case, SINCLAIR'S (1977) formula for predicting dietary protein from faecal protein correctly predicts the protein content of the Gilgil grasses from both zebra and cattle faecal protein (9.3% vs 9.9% and 9.8%, respectively), confirming that the assumption is valid in this case. The differences between the species on apparent nutrient extraction rates are, however, confusing. The two ungulate species are rather similar to each other on both energy and protein, as might be expected under the circumstances given that non-ruminant ungulates are more efficient than ruminants on poor quality herbage (JANIS, 1976). However, while

Grass-eating in Gelada

5

the two primates are less efficient than the ungulates on protein extraction (as would be expected), they appear to be slightly more efficient on energy extraction. This last result is somewhat difficult to interpret owing to the fact that there are a number o f possible confounding factors. The two most important are" (1) the fact that the various samples were obtained under different seasonal and habitat conditions (with consequential differences in baseline nutrient content and digestibility); and (2) the fact that baboons never feed entirely on grasses, but include a significant quantity of the more easily digested fruits and roots in their diet (see DUNBAR & DUNBAR, 1974). Nonetheless, it seems reasonable to conclude that the two ungulates are generally more efficient than the two primates at extracting proteins from their food. That baboons are as efficient as the gelada may reflect the dietary differences between them: even when grass is an important dietary item for them, baboons are never as exclusively graminivorous as the gelada. Since fruits (1) tend to be low in protein compared to grasses and (2) tend to have their nutrients in more easily assimilated forms, it may not be too surprising that baboons waste similar quantities of proteins to the gelada. The fact that fruits are relatively rich in energy compared to grasses (and have that energy in more easily assimilated forms) may also explain why baboons are apparently able to extract a higher proportion of the energy from their food than the two ungulate species. However, this cannot explain why the gelada should be equally efficient. The most likely explanation is that the apparent efficiency of the gelada relative to both cattle and zebra reflects habitat differences in the digestibility of the herbage. The digestibility of grasses is generally reduced by about 50% when the herbage stops growing and dessicates during the dry season (BRAUN, 1973). At high altitudes, ambient temperatures are generally lower and grasses dessicate less quickly and may therefore be more digestible for a greater proportion of the year. Given that the protein contents of the grasses in the two habitats do not appear to differ all that much, the extraction rates achieved by ungulates are impressive when compared to those for the two primate species. Very approximately, it appears that while the baboons are able to extract about 33 - 50% of the protein content of their forage, the zebra and cattle are able to extract around 80%. Even if the protein content of the forage at Ankober is closer to the higher values for Sankaber, it is clear that the gelada are significantly less efficient at protein extraction than either of the two ungulates. This greater efficiency on the part of the ungulates presumably reflects the fact that caecal and foregut fermentation makes proteins more accessible to them. This would also imply that the gelada do not do a significant amount of hindgut fermentation (see also IWAMOTO, 1979). Though strongly suggestive, however, it is clear that these data need verifying with analyses of nutrient inputs and outputs from populations of these species that live sympatrically. Despite this intrinsic weakness, these results have important implications for baboons' abilities to exploit a grass-based diet. DEMMENT and VAN SOEST (1985) found a significant difference in the body size distributions of East African ruminant and non-ruminant herbivores, with a critical threshold at around 10 kg body weight. They argue that there is a range of body weights between 10 - 600 kg at which ruminants are much more efficient at exploiting high fibre diets. Gelada lie within this range and clearly would face significant ecological competition from ruminants of similar body size, as well as finding it difficult to meet their daily nutrient requirements. This may explain why gelada are now confined to an area o f high altitude savannah/moorland where high rainfall and more equitable temperatures ensure that grasses are more digestible for a longer period of the year than is the case at lower altitudes. It may also explain why the gelada feed on grass blades only so long as

6

R . I . M . DUNBAR& U. BOSE

they remain green, preferring instead to dig for roots and rhizomes during the dry season once the grasses become dessicated and less digestible (DUNBAR, 1977, 1978; IWAMOTO& DUNBAR, 1983). These findings throw light on the as yet unresolved question as to why all the gelada's sister-species went extinct despite having been a major component of both the primate and herbivore biomasses on the eastern and southern African savannah grasslands during the Plio-Pleistocene (see JOLLY, 1972; ECK & JABLONSKI, 1987). Northeastern Africa, in particular, underwent a major change in climate at around two million years BP, associated with a shift in the typical vegetation from woodland to the dry thornscrub savannah that now dominates the area (BONNEFILLE, 1976). If this was associated with a more fibrous, less easily digested food intake, then it is likely that a gelada-like species would have had great difficulty meeting its daily nutritional requirements in such an environment.

Acknowledgments. R.D. is grateful to MIKEHARRISONfor collecting the gelada samples, to RICHARD DANSIE for permission to work on the Kekopey Ranch at Gilgil and to the Office of the President of Kenya for permission to carry out research in Kenya. We are much indebted to PETER DONACHIE for help with the nutritional analyses and to VASANTIAMIN for carrying out the second series of energy analyses. SIMONSTRICKLANDkindly advised us on nutritional matters.

REFERENCES BAUCHOP,T. & R. W. MARTUCC1,1968. Ruminant-like digestion of the langur monkey. Science, 161: 698 - 700. BONNEEILLE, R., 1976. Palynological evidence for an important change in the vegetation of the Omo basin between 2.5 and 2 million years ago. In: Earliest Man and the Environments in the Lake Rudolf Basin, Y. COPPENS, E C. HOWELL,G. LL. ISAAC,& R. LEAKEY(eds.), Chicago Univ. Press, Chicago, pp. 421-431. BRAUN, H. M. n. , 1973. Primary production in the Serengeti: purpose, methods, and some results of research. Ann. Univ. Abidjan, 6:171-188. DEMMENT, M. W. ~r P. J. VAN SOEST, 1985. A nutritional explanation for body-size pattern of ruminant and nonruminant herbivores. Amen. Naturalist, 125:641-672. DUNBAR, R. I. M., 1977. Feeding ecology of gelada baboons: a preliminary report. In: Primate Ecology, T. H. CLUTTON-BROCK(ed.), Academic Press, London, pp. 250-273. - - , 1978. Competition and niche separation in a high altitude herbivore community in Ethiopia. E. Afr. Wildl. J., 16: 183-199. - - , 1984. Theropithecines and hominids: contrasting solutions to the same ecological problems. J. Human Evol., 12: 647-658. - 8,; P. DUNBAR, 1974. Ecological relations and niche separation between sympatric terrestrial primates in Ethiopia. Folia Primatol., 21: 36-60. ECK, G. G. & N. G. JABLONSKI,1987. The skull of Theropithecus brumpti compared with those of other species of the genus Theropithecus. In: Les Faunes Plio-Pleistocenes de la Vallde de l'Omo (Ethiopie). 3. Cercopithecidae de la Formation de Shungura, G. G. ECK, N. G. JABLONSK1,& M. LEAKEY(eds.), Editions du CNRS, Paris, pp. 18-122. HARDING, R. S. O., 1975. Ranging patterns of a troop of baboons (Papio anubis) in Kenya. Folia PrimatoL, 25: 143-185. IWAMOTO,T., 1979. Feeding ecology. In: Ecological and Sociological Studies of Gelada Baboons, M. KAWAI(ed.), Karger, Basel, pp. 280-330. - & R. I. M. DUNBAR, 1983. Thermoregulation, habitat quality and the behavioural ecology of gelada baboons. J. Anim. Ecol., 52: 357-366. JANIS, C., 1976. The evolutionary strategy of the Equidae and the origins of rumen and caecal digestion. Evolution, 30: 757-774.

Grass-eating in Gelada

7

JOLLY, C. J., 1972. The classification and natural history of Theropithecus (Simopithecus) (ANDREWS, 1916), baboons of the African Plio-Pleistocene. Bull. Brit. Mus. Nat. Hist. (Geol.), 22:1 - 123. KAY, R. F., P. HOPPE, & G. M. O. MALOIY, 1976. Fermentative digestion of food in the colobus monkey, Colobus polykomos. Experientia, 32: 4 8 5 - 486. LEE-THORP, J. A., N. J. VAN DER MERWE, &; C. K. BRAIN, 1989. Isotopic evidence for dietary differences between two extinct baboon species from Swartkrans. J. Human Evol., 18: 183- 190. LESLIE, D. M. & E. E. STARKEY, 1985. Fecal indices to dietary quality of cervids in old-growth forests. J. Wildl. Mgmt., 49: 142-151. MILXON, K. & R. MCBEE, 1982. Structural carbohydrate digestion in a New World primate, Alouatta palliata GRAY. Comp. Biochem. Physiol., 74: 29-31. PHILLIPSON, A. T., 1977. Ruminant digestion. In: Duke's Physiology of Domestic Animals (9th ed.), M. J. SWENSON(ed.), Cornell Univ. Press, Ithaca, pp. 250-286. SINCLAIR, A. R. E., 1977. The African Buffalo: A Study of Resource Limitation of Populations. Chicago Univ. Press, London. VAN SOEST, P. J., 1982. Nutritional Ecology of the Ruminants. O & B Books, Corvallis, Oregon. STEVENS, C. E., 1988. Comparative physiology of the vertebrate digestive system. In: Comparative Nutrition, K. BLAXTER& I. MACDONALD(eds.), Libbey, London, pp. 2 1 - 36. SZALAY, E S. & E. DELSON, 1979. Evolutionary History of the Primates. Academic Press, New York.

--

Received September 15, 1989; Accepted May 30, 1990

Authors' Names and Address: R. I. M. DUNBARand UrPAULBOSE, Department of Anthropology, University

College London, Gower Street, London WC1E 6BT, England.