One hundred nineteen patients who met specific criteria for atypical depression completed six weeks of double-blind, randomly assigned treatment with phenelzine sulfate, imipramine hydrochloride, or placebo. The overall response rates were 71% with phenelzine, 50% with imipramine, and 28% with placebo. Phenelzine was widely superior to placebo and also showed superiority to imipramine. Phenelzine superiority appeared even greater after an additional six-week continuation phase. Imipramine was only moderately effective in this atypical depressive sample. Unexpectedly, the superiority of either phenelzine or imipramine to placebo was largely confined to patients in subsets of the study sample who were prospectively judged to also have a history of spontaneous panic attacks and/or show hysteroid dysphoric features. This is consonant with some but not other recent findings and requires replication. Overall, the concept of atypical depression as a subtype that is preferentially responsive to monoamine oxidase inhibitors is supported.