Low oral doses of melatonin raise serum melatonin concentrations to those normally occurring nocturnally and facilitate polysomnographically assessed sleep onset when given at different time points throughout the day, without altering mood or performance on the morning following treatment. In the present study, 12 young healthy volunteers, free of sleep disturbances, received 0.3 or 1.0 mg of melatonin or placebo at 2100 hours, 2-4 hours prior to their habitual bedtime. Polysomnographic recording of overnight sleep began at 2200 hours and continued until 0700 hours the following morning, when subjects were awakened. Sleep onset latency and latency to stage 2 sleep were significantly decreased as a result of melatonin treatment. Neither dose of melatonin significantly altered sleep architecture. Administration of the lower dose of melatonin (0.3 mg) at 2100 hours elevated serum melatonin to levels within the normal nocturnal range (113 +/- 13.5 pg/ml) at the time the sleep test was initiated. Neither melatonin dose caused "hangover effects", as assessed by self-reports or by mood and performance tests administered on the morning following treatment. These observations provide additional evidence that nocturnal melatonin secretion has a sleep-promoting function. They also indicate that an increase in serum melatonin concentrations, within the normal physiologic range, does not significantly alter sleep architecture in subjects with normal sleep who receive the treatment several hours prior to their habitual bedtime.