A detailed analysis of the results of a multi-centre clinical trial shows that, while the relapse rate following recovery from an operationally defined depressive illness was smaller among patients subsequently treated with either amitryptiline or lithium than with a placebo, there was no clinically significant difference between the prophylactic efficacy of the 2 antidepressants. An account is given of the relative adverse effects of the treatments, and the implications of the findings are discussed.