The comparative efficacy of propranolol and amitriptyline in the prophylaxis of migraine headache was studied in 30 patients in a double-blind, placebo-controlled, crossover design. Headache response to medication was measured monthly by compilation of headache scores derived from quantitative data recorded by patients in a daily diary; at each visit, Zung and Hamilton tests for depression and the Spielberger state test for anxiety were performed. In the absence of clinical toxicity at monthly visits, the decision to maintain the current dose or raise it was made by a computer, which compared current headache score with that of the previous month. Both drugs were superior to placebo. Neither drug was superior to the other. The effectiveness of neither drug correlated with a decrease in anxiety or depression demonstrated by psychological testing.