To assess whether lithium carbonate augments antidepressant effects of long-term antidepressant treatment in non-responding patients, 15 treatment-refractory patients were studied using a placebo-controlled, double-blind design. After at least 21 days of antidepressant drug therapy, and while continuing to receive the same daily dose of antidepressant drug, eight patients received lithium carbonate and seven received placebo. In comparison with placebo, lithium carbonate produced a small but statistically significant improvement in the mean daily nursing ratings of depression during the first two days of treatment. The beneficial effects of lithium carbonate were more variable during the next four days, but by the seventh through 12th day of the trial, the drug produced a significant and clinically meaningful improvement. When the seven placebo-treated patients received active lithium carbonate on the 13th day of the study, their rate of improvement was similar to that of the eight patients who had received active lithium carbonate initially. This augmentation of the anti-depressant effect was seen in patients treated with desipramine hydrochloride, amitriptyline hydrochloride, or mianserin hydrochloride. Although in five of the 15 patients, the improvement appeared as early as 24 to 48 hours after the first lithium carbonate dose, the remaining patients did not show a clear improvement until approximately five to eight days later. We concluded that lithium carbonate does augment the antidepressant effect when added to the long-term antidepressant treatment of nonresponding patients.