318 patients satisfying the Ad Hoc Committee's criteria for common or classical migraine were entered into an 8 week single-blind placebo recording phase to establish, by diary cards, the frequency and severity of their attacks. 176 patients completed this and had records indicating 4-8 episodes in the 8 week period, with sufficient severity to reduce activity and/or work; these patients were randomized by a predetermined code, into three double-blinded groups: naproxen sodium 550 mg bid (60 patients), pizotyline 0.5 mg tid (59 patients), or placebo (57 patients). The patients were followed at monthly intervals for 12 weeks, with 25 dropping out (3 on naproxen sodium, and 2 each on pizotyline and placebo because of "side effects;" the remaining 18 because of noncompliance or reasons unrelated to therapy). Approximately 25% of patients in each of the 3 groups complained of side effects. Statistical analysis showed that both naproxen sodium and pizotyline were better than placebo, and of overall equivalent (i.e. equal) efficacy in the prophylaxis of migraine. In some respects, naproxen sodium was slightly more effective than pizotyline in the first month of treatment.