Twelve volunteer poor sleepers of mean age 59 years took placebo on one night, promethazine 20 mg on one night and promethazine 40 mg on one night, in a double-blind balanced order study. Sleep in the EEG laboratory was increased by nearly 1 h after either dose of promethazine, and sleep interruptions were reduced. Slow-wave sleep was unaffected, but the larger dose reduced the percentage of sleep spent as REM sleep. Sleep was improved subjectively by both doses of promethazine which appears to be an effective hypnotic.