Objectives: The study examined whether patients who responded to an antidepressant in one episode of depression responded to the same agent in a subsequent episode. It also sought to determine whether in a subsequent episode clinicians prescribed an antidepressant that bad previously been effective. Methods: Charts of inpatients with a diagnosis of major depression who were consecutively admitted to general hospital psychiatric wards were screened to determine whether the patients met the inclusion criteria of recurrent major depression and had responded to antidepressant therapy in a previous episode. Data were collectedon type of treatment and treatment response for each episode. Results: Fifty-nine patients met the inclusion criteria. Thirty-five of the patients (59.3 percent) were prescribed the same antidepressant in both the current ad previous episodes. Of these, 20 patients responded and 15 did not. Of the 24 patients who did not receive the same antidepressant, 14 were attended by the same dinician in both treatment episodes. Nineteen of the 24 patients responded to a new antidepressant, and five did not. Conclusions: Patients with recurrent depression are often prescribed antidepressants other than those that were effective in previous treatment episodes. Use of a previously effective antidepressant may not increase the treatment response rate in a subsequent depressive episode.