As a partial test of whether the same or different patients benefit from cognitive therapy and tricyclic antidepressant agents, depressed outpatients first received cognitive therapy, then nonresponders were treated with either imipramine or placebo. If the two treatments were effective for the same subgroup of patients, imipramine should not be more effective than placebo because potential responders should already have been removed by treatment with cognitive therapy. Alternatively, if cognitive therapy and imipramine are effective for different subtypes of depressive disorder, then imipramine ought to be more effective than placebo for patients failing to benefit from cognitive therapy because some potential imipramine failures would already have been removed. Thirty-six depressed outpatients were treated with weekly cognitive therapy for 16 weeks with 17 (47%) responding. Nonresponders were then randomly assigned to imipramine or placebo for 6 weeks to a maximum dose of 300 mg of imipramine per day. Of 12 patients completing the double-blind medication trial, all 5 assigned to imipramine had a clear-cut response, whereas none of the other seven benefited from placebo (χ2= 12.00;p= 0.001). Although the numbers are small, these results suggest rejection of the hypothesis that imipramine is effective for the same subpopulation of depressed patients as is cognitive therapy.