Diazepam 5 mg, promethazine 25 mg, propiomazine 25 mg, and placebo as sleeping aids were studied in 40 elderly inpatients. The drugs were administered in a random order, double-blind, on 21 nights. All active drugs were found effective in the mentally normal elderly, but in the psychogeriatric group of 20 patients, only propiomazine very significantly reduced the number of nocturnal awakenings and only diazepam almost significantly shortened the initial sleep latency, although duration of sleep was prolonged in both groups. There was neither loss of efficacy of the trial drugs nor rebound on withdrawal. Psychomotor skills and serum prolactin levels remained unaffected.