The main purpose of these studies is to explore the mechanism(s) by which bright light exerts an antidepressant effect in patients with winter seasonal affective disorder (SAD). It has been difficult to rule out a placebo effect because of the difficulty in designing an adequate placebo control. There are several problems with the use of dim light, the most commonly used placebo, primarily that patients usually expect greater benefit from bright light. To avoid this problem we have developed an unique 'environmental' placebo. Patients will be tested in five-week parallel groups design with three groups, morning light, evening light and placebo. Thus we will be able to compare the antidepressant effect of morning light vs. evening light, morning light vs. placebo, and evening light vs. placebo. We will also evaluate the phase-shift theory that the antidepressant mechanism of bright light is due to shifting circadian rhythms. Half of the SAD subjects in each of the three groups will have their circadian rhythm of body temperature continuously measured for the five weeks. Age and sex matched normals will serve as controls. We will also study a group of subjects who have the core symptoms of SAD, decreased energy, increased sleep, and increased appetite, but who do not meet criteria for major depression. A similar sample has been termed subsyndromal SAD. Epidemiological studies suggest that a large group of people in temperate latitudes fall into this category. We will study their response to bright light with the same design mentioned above. As a further means of exploring mechanisms the fall onset and spring offset of the various symptoms will be charted over two years to determine sequence and to document the seasonal nature of the syndrome.