The author reviews the studies of lithium as an antidepressant, evaluates the evidence that there might be specific subgroup of patients for whom it is effective, and reports a study he and his associates conducted that found lithium to be effective for 13 of 21 depressed patients. He concludes that there is convincing, although not conclusive, evidence for an antidepressant effect of lithium and that only by identifying the subgroup of patients for whom it is effective can the continuing uncertainty surrounding lithium's role in affective disorder be resolved. He also calls for a reevaluation of the relationship between mania and depression.