A meta-analysis of the controlled clinical trials comparing the efficacy of combined lithium-imipramine therapy with each drug alone in the prevention of relapse in patients with unipolar depressive illness shows the combination to be superior to lithium alone (for any relapse, p less than 0.05; for specifically depressive relapse, p less than 0.025) and imipramine alone (for any relapse, p less than 0.025; for specifically depressive relapse, p less than 0.05). The issue of statistical power in studies of this kind is discussed. A single case of unipolar depressive illness is described, in which combined lithium-tranylcypromine therapy was superior to either drug alone. A combination of lithium and an antidepressant should be considered for any patient suffering from this condition who fails to remain well on either drug alone.