MDD and AD, while differing greatly in clinical course and pathophysiology, have important interactions. MDD in late-life appears to be a risk factor for the development of AD for reasons that are not apparent. The prevalence of AD in MDD is probably higher than in the general elderly population, but MDD does not appear to be a frequent complication of AD, However, MDD in AD patients worsens both cognition and function in ADL. MDD appears to be overdiagnosed in AD, especially when the diagnosis is based on caregiver reports. Much of what are thought to be symptoms of MDD are, in fact, direct consequences of AD, such as difficulty with initiating activities and sustaining attention. When MDD occurs in AD, it is treatable by conventional pharmacological means with concomitant simplification of environmental demands and development of a structured routine. The major confound in the scientific investigation of MDD in AD is differentiating between symptoms engendered directly by AD and symptoms of MDD. The omission in DSM-IV of depressed mood as a requisite criterion for the diagnosis of MDD will add still more difficulty. The authors propose an alternative to DSM-IV criteria for MDD for the diagnosis of MDD in persons with dementia. These alternative criteria shorten the period of time during which symptoms must manifest to 1 week, make depressed mood a requirement, allow for lack of initiative, and substitute irritability for sleep disturbance. Finally, given the relatively benign side-effects profile of modern antidepressant agents, it is reasonable that persons with AD whose behavior suggests that they may be depressed be given a trial of antidepressant medication.