.---The two species of African oxpeckers (Buphagus africanus and B. erythrorhynchus) have been reported to be un- differentiated in food choice (Artwell 1966, Puku 4: 17; Olivier and Laurie 1974, Auk 91: 169). This conclusion is supported by similarity in their stomach con- tents (Moreau 1933, Bull. Entomol. Res. 24: 325; van Someten 1951, E. African Agr. J. 17: 1) and by their mutual and sometimes simultaneous occurrence on cer- tain species of large mammals where they feed on ticks, flies, and tissue from sores (Attwell ibid., Olivier and Laurie ibid.). The importance of differential choice of foraging substrates (i.e. species of large mammals) in reducing interspecific ecological overlap between these two species has not been identified. TABLE 1 OCCURRENCES 10J  YELLOW-BILLED OXPECKERS (BUPHAGUS AFRICANUS) AND RED-BILLED OXPECKERS (B. ERYTHRORHYNCHUS) ON LARGE MA2kœ3/FALS Okavango Swamp'-' Zambia  Mammals B.a. B.e. B.a. B.e. Naked or fur sparse, parasites eaten: Black rhinoceros (Diceros bicornis) -- -- 12 8 Wart hog (Phacochoerus aethiopicus) 0 3 3 4 Buffalo (Syncerus caller) 5 2 35 7 Subtotal 4 5 5 50 19 Fur moderately dense, parasites eaten: Burchell's zebra (Equus burchelli) 0 1 4 7 Giraffe (Giraffa camelopardalis) 0 5 2 2 Impala (Aepyceros melampus) 0 7 3 34 Roan (Hippotragus equinus) -- -- 16 16 Sable (Hippotragus niger) -- -- 2 9 Blue wildebeest (Connochaetes taurinus) 0 1 3 0 Kudu (Tragelaphus strepsiceros) 0 4 9 14 Eland (Taurotragus oryx) -- -- 6 11 Subtotal 0 18 45 93 Naked, tissue eaten: Hippopotamus (Hippopotamus amphibius) -- -- 10 36 Subtotal 0 0 10 36 Separate observations of one or more oxpeckers on an individual or herd of a mammal species. This study. Z Attwell ibid., combined data from Kafue National Park and Luangwa Valley. Subtotals used in statistical tests. During June and July 1973 I recorded the substrate choices of oxpeckers in the vicinity of Chief's Island, Okavango Swamp, northern Botswana. These data demonstrate a significant difference in substrate choice between the two oxpecker species (Table 1, P  0.001, Fisher Exact Probability test). The Yellow-billed Oxpecker (B. africanus) occurred exclusively on buffalo, and the Red-billed Ox- pecker (B. erythrorhynchus) usually foraged on other species. B. erythrorynchus was in smaller groups (1-23, mean 4.9 birds) than B. afrlcanus (1-100, mean 32.0 birds). Thus, the relative abundance of the two species in the swamp was less disproportionate than the frequencies of sightings indicate. Attwell's (ibid.) data from Zambia show a similar differentiation in substrate choices (Table 1, P  0.001, Chi-square test, df = 2) and preference by B. africanus for buffalo and other sparsely furred or naked substrates. Morphologically, oxpeckers differ in bill structure: B. africanus has a broader, heavier bill; B. erythrorhynchus a narrower, more scissorslike bill (Attwell ibid.). The correlation between broad bill form and the choice of sparsely furred or naked substrates (e.g. buffalo and rhinoceros) in B. africanus and narrow bill form and the preferential use of furred substrates (e.g. impala, kudu, and giraffe) in B. erythrorhynchus suggests that bill differentiation is related to fur density. Attwell (ibid.) has suggested that lateral compression of the bill may facilitate grasping or capturing parasites and flies in fur. Both oxpecker species avoid substrates with dense, woolly fur (e.g. reedbuck Redunca arundinum, waterbuck Kobus elllpsi- prymus, and lechwe Kobus leche). The relatively great affinity of B. erythrorhyn- chus for hippopotamuses, which lack fur, would seem to contradict this relation- ship, though all foraging on hippopotamuses is on tissue from open wounds (Olivier and Laurie ibid.). The importance of bill shape in taking this type of food is unknown, but a narrow bill could possibly be advantageous. Artwell (ibid.) considers the preferences of B. africanus for buffalo and of B. erythrorhyn- chus for domestic cattle as paradoxical on the basis of the close taxonomic relation- ship between the substrates, but these choices are predicted by the suggested relationship between bill shape and the density of substrate fur. I thank W. Hamilton, J. Hunt, and J. Wittenberger for helpful criticisms of the manuscript. Fieldwork was supported by a NSF Grant (GB 28533) to the University of California, Davis.--WLLA H. Busxmx, Division of Environmental Studies, University of California, Davis, California 95616. Present address: Earlham College, Richmond, Indiana 47374. Accepted 18 Jun. 74.