In antidepressant drug trials, the randomized treatment phase is usually preceded by a single-blind placebo period or "washout." Ninety-four depressed patients who improved during this 10-day placebo period constitute our study focus. Analysis of baseline and postplacebo measures showed that the 10-day placebo responders in our sample were convincingly depressed at baseline and improved significantly after placebo washout. This group of patients differed from 6-week placebo responders in our randomized trials in being more mildly ill, being more chronic, containing fewer cases of primary depression, and having fewer illness precipitants. They differed from placebo nonresponders largely in manifesting milder illness symptoms across the range of psychopathology. The proportion of placebo washout responders declined in the winter months.