One hundred and eight healthy controls and 178 patients with a major depressive disorder according to DSM-III were investigated in the sleep laboratory after a 7-day drug wash-out period. Subsamples of 36 healthy controls and 56 patients additionally took part in the cholinergic rapid eye movement (REM) sleep induction test with RS 86. Data analysis revealed that age exerted powerful influences on sleep in control subjects and depressed patients. Sleep efficiency and amount of slow wave sleep (SWS) decreased with age, whereas the number of awakenings, early morning awakening, and amounts of wake time and stage 1 increased with age. REM latency was negatively correlated with age only in the group of patients with a major depression. Statistical analysis revealed group differences for almost all parameters of sleep continuity with disturbed indices in the depressed group. Differences in SWS were not detected. REM latency and REM density were altered in depression compared to healthy subjects. Sex differences existed for the amounts of stage 1 and SWS. The cholinergic REM induction test resulted in a significantly more pronounced induction of REM sleep in depressed patients compared with healthy controls, provoking sleep onset REM periods as well in those depressed patients showing baseline REM latencies in the normal range. Depressed patients with or without melancholia (according to DSM-III) did not differ from each other, either concerning baseline sleep or with respect to the results of the cholinergic REM induction test. The results stress the importance of age when comparing sleep patterns of healthy controls with those of depressed patients. Furthermore they underline the usefulness of the cholinergic REM induction test for differentiating depressed patients from healthy controls and support the reciprocal interaction model of nonREM-REM regulation and the cholinergic-aminergic imbalance hypothesis of affective disorders.