The authors investigated lithium prophylaxis of depression in bipolar I (N=35), bipolar II (N=18), and unipolar I (N=28) outpatients. Depression indices included frequency of depressive episodes, three measures of severity, dropouts due to depression, and time in study. There were significantly fewer episodes and dropouts due to depression in the lithium-treated than in the placebo-treated bipolar I patients. Lithium was superior to placebo on all indices in the bipolar II group (although the sample size did not permit statistical analysis) and on 3 indices in the unipolar group. The authors believe that lithium prophylaxis of depression has been demonstrated in these three subtypes but note the need for further research particularly in the use of tricyclic antidepressants.