Thirty depressed patients were deprived of sleep during the second half of the night after they were characterized according to their Research Diagnostic Criteria diagnosis, severity of depression, current medication and dexamethasone suppression test status. After applying partial sleep deprivation (PSD) treatment three times within 1 week, we observed in 17 patients an improvement of their initial depressive symptomatology of 50% or more. No pretreatment variable except the degree of improvement after the first PSD was predictive for therapeutic effects of the entire PSD series. Beneficial effects of PSDs were not specific for the endogenous type of depression. A trend was found indicating that concurrent psychotropic drug treatment enhances the beneficial effect of serial PSDs.