Sixty patients who met Research Diagnostic Criteria for major, intermittent, or minor depressive disorder and had reactive mood without atypical symptoms were treated with imipramine hydrochloride, phenelzine sulfate, or a placebo. These patients, referred to as simple mood reactive depressives, were contrasted with previously published data from 180 atypical depressives. Atypical depressives had the presence of at least one vegetative atypical sign (hypersomnia, hyperphagia, leaden feeling, or rejection sensitivity) but were otherwise indistinguishable from simple mood reactive depressives. In contrast to the atypical depressives for whom phenelzine was effective and imipramine was relatively ineffective, both medications were equivalently good in simple mood reactive depressives. Since all groups did poorly when given a placebo and well when given phenelzine, the salient feature of atypical symptoms may be that they predict poor response to imipramine. Since the difference between imipramine and placebo depends on the diagnostic group, pharmacologic dissection suggests that atypical symptoms in patients with nonautonomous mood may delineate a qualitatively distinct subgroup.