Termination of early pregnancy, by vaginal administration of prostaglandin analogues, one to three weeks after the first missed menstrual period, has advantages and disadvantages in comparison with vacuum aspiration. Some of these may be reduced if the patient is treated earlier. In the present study the effect and safety of one vaginal administration of 2.5 to 3 mg 15-methyl-PGF2 alpha methyl ester around the expected time of menstruation was evaluated in 16 women exposed to the risk of pregnancy. The overall number of treatment cycles was 35 and pregnancy was confirmed by plasma beta-HCG in eight. The treatment resulted in bleeding in all the pregnant cycles while in the nonpregnant ones it only provoked spotting and bleeding did not begin until the expected time of menstruation. Treatment with 2.5 mg 15-methyl-PGF2 alpha methyl ester resulted in complete abortion in one of three women. If the dose was increased to 3 mg all five treated women aborted. In nonpregnant patients no changes in the levels of estradiol-17 beta or progesterone at any time during the 24-hour observation period were found. Serum cortisol and prolactin but not TSH levels started to increase two hours after the start of treatment and reached a maximum after five hours. The increase coincided with the onset of uterine pain. Ovulatory cycles as judged from basal body temperature occurred in the first menstrual cycle following treatment in all nonpregnant patients. Although possible to use as a "once a month treatment" it seems preferable since the dose is the same, to postpone treatment until menstruation is delayed for a week or more.