230Th and 231Pa dates were obtained on 10 bone samples from Quaternary deposits near Medicine Hat, Alberta. The youngest date so obtained (9500 ± 1500 years) agrees reasonably well with a radiocarbon date (11 200 ± 200 years) from the same site. In addition, where samples were collected from more than one bed at the same bluff the results are in the correct chronological order. A Sangamon bed containing chipped stones, thought to be artifacts, gave a date of 72 000 ± 6000 years, which is acceptable, though at the younger end of the previously estimated age range of between 200 000 and 70 000 years for the bed.In general, however, the results reveal large discrepancies with the estimated ages of the beds that yielded the bones. Although several causes for the discrepancies are possible, in two or three instances the uranium-series dates themselves appear to be incorrect, and the other dates must be treated with caution.