We included 43 patients with migraine without aura in a triple-blind, placebo- and dose-controlled, crossover study of the prophylactic effect of slow-release sodium valproate; 34 patients completed the trial. The number of days with migraine was 3.5 per 4 weeks during treatment with sodium valproate and 6.1 during placebo (p = 0.002). The severity and duration of the migraine attacks that did occur were not affected by sodium valproate when compared with placebo. Fifty percent of the patients were responders, ie, their initial migraine frequency was reduced to 50% or less during sodium valproate as compared with 18% during placebo. The number of responders increased during the trial to 65% in the last 4 weeks of the active treatment period. There were no serious side effects requiring withdrawal of patients from the study. We conclude that sodium valproate is an effective and well-tolerated prophylactic medication for migraine without aura.