The discovery of Australopithecus afarensis has led to new interpretations of hominid phylogeny, some of which reject A. africanus as an ancestor of Homo. Analysis of buccolingual tooth crown dimensions in australopithecines and Homo species by Johanson and White (Science 202:321-330, 1979) revealed that the South African gracile australopithecines are intermediate in size between Laetoli/hadar hominids and South African robust hominids. Homo, on the other hand, displays dimensions similar to those of A. afarensis and smaller than those of other australopithecines. These authors conclude, therefore, that A. africanus is derived in the direction of A. robustus and is not an ancestor of the Homo clade. However, there is a considerable time gap (ca. 800,000 years) between the Laetoli/Hadar specimens and the earliest Homo specimens; "gracile" hominids from Omo fit into this chronological gap and are from the same geographic area. Because the early specimens at Omo have been designated A. afarensis and the later specimens classified as Homo habilis, Omo offers a unique opportunity to test hypotheses concerning hominid evolution, especially regarding the phylogenetic status of A. africanus. Comparisons of mean cheek teeth breadths disclosed the significant (P less than or equal to 0.05) differences between the Omo sample and the Laetoli/Hadar fossils (P4, M2, and M3), the Homo fossils (P3, P4, M1, M2, and M1), and A. africanus (M3). Of the several possible interpretations of these data, it appears that the high degree of similarity between the Omo sample and the South African gracile australopithecine material warrants considering the two as geographical variants of A. africanus. The geographic, chronologic, and metric attributes of the Omo sample argue for its lineal affinity with A. afarensis and Homo. In conclusion, a consideration of hominid postcanine dental metrics provides no basis for removing A. africanus from the ancestry of the Homo lineage.