The effects of 400-600 mg trazodone on the sleep patterns of ten depressed in-patients treated for 5 weeks were studied during the initial (days 1-3) and terminal (days 26-28) treatment periods. The sleep parameters were compared to those obtained from three sleep recordings performed just prior to the initiation of the treatment and after 2 adaptation nights at the end of a 2-week drug-free period. At the same time, the clinical evolution of patients was evaluated weekly using MADRS and Hamilton-Anxiety scales for anxiety-depression symptomatology and Spiegel and Norris sleep scales. Weekly blood samples were collected to measure plasma levels of trazodone and, at the end of the study, the elimination half-life at steady state was calculated by repeated measurements of plasma levels. Clinical improvement, as assessed by a reduction of more than 60% in MADRS scale scores, was accompanied by evidence of the definitely beneficial effects of trazodone on the disturbed sleep of these depressed patients. From the beginning of treatment, there was a hypnotic-like effect (increase in total duration of sleep and stage II, decrease in sleep latency and intrasleep awakenings). In addition, records at the end of the study showed an increase in delta sleep and an increase in REM latency, an effect classically associated with an antidepressant action. These particularly valuable effects of trazodone on sleep would suggest that this drug should especially be given in cases of depression with major insomnia.