A study was carried out in which 135 mildly or moderately depressed outpatients were randomly allocated to one of five groups receiving six weeks' treatment weith antidepressant drugs. The groups received a tricyclic antidepressant (trimipramine; mean dose 106 mg at night) or a monoamine oxidase inhibitor (MAOI) (phenelzine or isocarboxazid; mean doses 45 and 32 mg/day respectively), or a combination of the two (phenelzine plus trimipramine or isocarboxazid plus trimipramine). Various scales were used to measure depression before and at one, three, and six weeks of treatment, and results were assessed blindly. The tricyclic antidepressant was found to be consistently superior to the MAOIs and the combined treatments. Some differential indicators of response to the various antidepressants were found--for example, patients with initial complaints of dizziness, suicidal ideas, irritability, and insomnia and a longer duration of illness were more likely to respond to trimipramine--but these were of only modest significance. Side effects were not troublesome in any group. It is concluded that neither MAOIs nor MAOIs combined with tricyclic antidepressants are the treatment of first choice in unselected outpatients with mild or moderate depression.