Perimenopause-related mood disorders cause significant distress to a potentially large number of women. Recent studies have reported the therapeutic benefits of estradiol in women with these mood disorders; however, a relevant pathophysiologic role of declining estradiol secretion during the perimenopause has not been demonstrated. In this protocol we wish to investigate the effects of acute withdrawal of estradiol on mood under placebo controlled conditions. Thus, mood and behavior symptoms may be precipitated by the experimental conditions of this protocol. This protocol will address the following hypothesis: women with a past history of perimenopause-related depression but not women without such a history will experience a recurrence of mood and behavioral symptoms during acute estradiol withdrawal but not during continued estradiol administration. The nature of the relationship between the declining secretion of estradiol and mood in perimenopausal depressed women will be examined as follows: Peri- and postmenopausal women reporting the onset of depression during the perimenopause and who report remission of depressive symptoms on estrogen therapy (ET) will be withdrawn from ET under blinded and placebo-controlled conditions. We will recruit as a comparison group asymptomatic women on hormone replacement and without a history of perimenopause-related depression. During a three week baseline phase, all women will be switched from their current form of hormone replacement therapy to estradiol and will complete symptom ratings to confirm the absence of mood symptoms prior to entry into the study. After the screening, all women will be randomized to receive either estradiol or placebo for an additional three weeks. Comparison of mood ratings during these contrasting treatment conditions will allow us to examine the specific role of estrogen withdrawal in depression that is responsive to ET.