Long renowned for their clues to the presence and predatory abilities of North America's earliest humans, Southern Plains sites with bison remains have provided vertebrate paleontologists with food for thought for nearly a century. Bison skulls from Texas, New Mexico, and Kansas figured prominently in early taxonomic schemes, ideas on bison evolution, and the chronology of geological deposits before radiocarbon dating. Now, in addition toinnumerable finds of isolated cranial parts, bison remains are reported for over 30 non archaeological sites from southwestern Kansas to southern Texas. A radiocarbon-dated chronology is slowly developing for these paleontological finds, and, on this basis, bisona remains appear to be nonexistent in deposits older than 50,000 yearsBP. Moreover, most localities older than 20,000 years BP yield more remains of herbivores other than bison. Although several competing models ofNorthAmerican bison evolution are currently in vogue, Southern Plains finds demonstrate that serious problems hinder full acceptance of any model. These problems include small sample sizes, poor understanding of age and sex differences within nominal species, inadequate dating of most faunal localities, and (most of all) identification errors resulting from misinterpretation of nomenclature. As a result, many identifications of bison species published in the last 20 years are suspect.