In several studies we have found that treatment with bright environmental light, capable of suppressing human melatonin, reverses the winter depressive symptoms of patients with seasonal affective disorder (SAD), whereas light too dim to suppress human melatonin is therapeutically ineffective. This finding, as well as the central importance of melatonin as a hormonal mediator of photoperiodic changes on seasonal rhythms in animals, led us to test the hypothesis that melatonin mediates the effects of shortening days on the winter symptoms of SAD and that the modification of melatonin secretion by bright light mediates its antidepressant effects. We partially reversed the antidepressant effects of phototherapy in 8 SAD patients by oral melatonin administration, but in another study of 19 SAD patients we failed to find any therapeutic difference between the beta-adrenergic blocker, atenolol, which inhibits melatonin secretion, and placebo. In a third study of 7 SAD patients we showed that the anti-depressant effects of phototherapy were not photoperiodic and appeared to be independent of melatonin suppression. There is some preliminary evidence that melatonin secretion may be abnormal in SAD. We conclude that while melatonin may play some role in the symptoms of SAD and the effects of phototherapy, it cannot by itself account for these phenomena.