The relation between self-reported physical activity and depressive symptoms was analyzed for 1,900 healthy subjects aged 25-77 years in the Epidemiologic Follow-up Study (1982-1984) to the first National Health and Nutrition Examination Survey (NHANES I). Depressive symptomatology as measured by the Center for Epidemiologic Studies Depression Scale (CES-D) was examined by sex and race in relation to recreational physical activity and physical activity apart from recreation, controlling for age, education, income, employment status, and chronic conditions. Little or no recreational physical activity and little or no physical activity apart from recreation were cross-sectionally associated with depressive symptoms in whites and in blacks. After exclusion of those with depressive symptoms at baseline, recreational physical activity was an independent predictor of depressive symptoms an average of eight years later in white women. The adjusted odds of depressive symptoms at follow-up were approximately 2 for women with little or no recreational physical activity compared with women with much or moderate recreational physical activity (95% confidence interval 1.1-3.2). These findings are the first indication from a prospective study of a large community sample that physical inactivity may be a risk factor for depressive symptoms.