Recent studies have shown that melatonin is both hypnotic and hypothermic at physiological levels. Indeed, the hypnotic effect may be mediated via the hypothermic action. If this is the case, it is important to explore the dose-response relationships for the thermoregulatory effects of melatonin. Four groups of eight healthy adults (n = 32), aged between 18 and 38, each underwent two 12-hr bedrest protocols in which core body temperature (CT) and plasma melatonin levels were measured concurrently between 0800 and 2000 hr. For each group, subjects ingested either sucrose placebo or a 0.1, 0.5, 1.0, or 5.0 mg melatonin capsule at 1600 hr in a double-blind counterbalanced cross-over design. Melatonin was absorbed rapidly, with peak levels being reached after 1 hr at all dose levels. Mean peak plasma melatonin levels increased from physiological to pharmacological levels in a dose-dependent manner. Elimination for all dose levels was rapid, with mean plasma half-lives between 33 and 47 min. At the lower doses the mean drop in CT was between 0.05 and 0.15 degrees C and took between 2 and 3 hr. At the higher doses (1.0 and 5.0 mg), CT fell by 0.25-0.3 degrees C within 30-60 min following ingestion and at the highest dose (5 mg) remained suppressed for the duration of the study. While the magnitude and duration of the drop increased in what appeared to be a dose-dependent manner, it is unlikely that this relationship reflects a simple dose-response curve. There was considerable variability in plasma profiles following administration, particularly at the two lowest doses (0.1 and 0.5 mg). The lower mean drop in CT probably reflects the lower proportion of subjects achieving physiological plasma levels, and therefore a hypothermic effect, at the two lowest doses. If melatonin is to be used to improve sleep onset and maintenance by lowering CT, doses between 1.0 and 5.0 mg appear to be the lowest that produce a consistent drop in CT across all subjects.