Thirty-two patients entered a simple-blind, cross-over study on the action of tocopherol in osteoarthritis; only 3 did not complete the course. Each patient was randomly assigned either to the tocopherol group (600 mg/day for 10 days) or to the placebo group. After 10 days the groups were transposed. The analgesic and other possible effects of tocopherol vs. placebo were assessed by the patients' daily records, by the physician's personal examination and interview, and by observations on the use of an additionally permitted analgesic (pro re nata). In 52 percent of the 29 patients who completed the study of good tocopherol analgesic effect was noted, but only 4 percent of those receiving placebo reported a similar effect. The difference was statistically significant. Further large-scale assessement of the influence of tocopherol in osteoarthritis would seem to be justified.