The phylogenetic relationships of early Pleistocene Homo crania from the South African sites of Swartkrans and Sterkfontein were investigated through cladistic analyses of 99 morphological characters. The Swartkrans Member 1 specimen SK 847 and the Stw 53 cranium from Sterkfontein Member 5A were treated as separate operational taxonomic units (OTUs), distinct from the three species of early Homo-H. erectus, H. habilis, and H. rudolfensis-that are recognized from the Plio-Pleistocene deposits of East Africa. The cladistic analyses differed in the treatment of the South African OTUs (separate Swartkrans and Sterkfontein OTUs vs. a single Swartkrans+Sterkfontein OTU). PAUP 4.0 was used to construct cladograms and address hypotheses about relationships. In the analysis that treated the South African specimens as a single OTU, the position of that OTU was stable as a separate branch on the Homo clade between H. rudolfensis and [H. habilis+(H. erectus+H. sapiens)]. When SK 847 and Stw 53 were treated as separate OTUs, the majority of most parsimonious trees indicated that they were positioned in similar positions as the combined South African Homo OTU; that is, as separate branches between H. rudolfensis and [H. habilis+(H. erectus+H. sapiens)], with the Swartkrans OTU generally occupying a more derived position. The position of the Sterkfontein OTU was more stable than that of the Swartkrans OTU, which was found in several other positions among the minimum length trees. Running the analyses with only those characters preserved by SK 847 and Stw 53 resulted in similar topologies for minimum length trees, although the positions of Stw 53, SK 847, and H. habilis exchanged places in some trees. In no case was an exclusive sister relationship between either South African OTU and a particular species of Homo supported statistically. Both South African OTUs differ from H. habilis in the fewest number of cladistic characters.