Patients with winter depression (seasonal affective disorder) respond beneficially to sleep deprivation and bright light, but the mechanisms of these responses remain unknown. The study was designed to test whether afternoon/evening melatonin can prevent further relapse after sleep deprivation (presumably due to a pharmacologically induced advance shift of circadian phase). Compared to phase advancing by alteration of sleep-wake schedule or by bright light exposure, the melatonin intake is a more tolerated treatment procedure, and it provides a possibility of blind comparison between chronotherapeutic and placebo treatments. The depression was scored in 16 female patients with winter depression and 17 age-matched female controls before and after total night sleep deprivation and after subsequent 6-day administration of melatonin (0.5 mg) or placebo under double blind conditions. The melatonin intake was scheduled at 17:00 in order to produce a phase advance of circadian rhythms. Sleep deprivation resulted in 38% reduction of depression score in patients, but it did not reduce depression score in controls. After the subsequent treatment with placebo or melatonin slight but significant improvement of mood was found in controls. These treatments also stabilized the antidepressant response to sleep deprivation in patients. However, neither differential effect of melatonin and placebo on depression score nor alteration of habitual sleep timing was found in patients and controls. We conclude that the study results do not provide evidence for antidepressant potential of melatonin in patients with winter depression under realistic clinical conditions. The finding of stabilization of mood in patients with placebo suggests the contribution of psychological factors to therapeutic action of this and other innovative treatments for winter depression. The study was supported by the Russian fund of the Society for Light Treatment and Biological Rhythms.