26 women aged 15-45 with severe, primary dysmenorrhea were treated with naproxen (NAPROSYN, SYNTEX) and placebo during 2 x 2 consecutive menstrual cycles in a randomized, double-blind crossover study. The dosage of naproxen was 500 mg (2 tablets) initially, followed by 250 mg as needed, with a maximum of 1250 mg daily. In most cases medication started at the first sign of menstrual distress. 80 per cent of the women preferred naproxen to placebo. The number of tablets taken during each menstruation fell from a mean of 17.8 in the placebo period to 5.1 in the naproxen period. Likewise, additional analgesics fell from 7.1 to 1.6 and hours of bed rest from 16.4 to 1.2. Total number of days of sick leave per two menstruations decreased from 40 to 7. These differences are statistically significant (P < 0.001). The side effects were mild. CNS or gastrointestinal side effects were not seen. Naproxen changed the amount of bleeding in 12 and delayed bleeding in three. Two developed acne, which however gradually diminished during the next five bleeding periods treated with naproxen. The influence of prostaglandin synthetase inhibitors on the ovarian production of steroids is discussed.