We measured serum progesterone (P) levels after administration of 400 mg P vaginal suppositories to women during the luteal phase of the menstrual cycle. Blood samples were obtained before suppository insertion and at five intervals up to 8 hours after insertion. On the first day of treatment with P suppositories, there was a substantial elevation in serum P above baseline after insertion. However, on subsequent days of administration a smaller increment in serum P was observed. In 4 women studied on days 1 and 8 of the same treatment cycle in the luteal phase, a smaller rise in serum P following suppository administration on day 8, compared with day 1, was found. Overall, a highly significant negative correlation between change in serum P from baseline and duration of vaginal suppository treatment was found. This observation does not appear to be related to the achievement of a pharmacokinetic steady state. Possible mechanisms for this observation are discussed.