Two waves of data from a community-based study (Alameda County Study, 1994-1995) were used to investigate the association between obesity and depression. Depression was measured with 12 items covering Diagnostic and Statistical Manual of Mental Disorders: DSM-IV diagnostic criteria for major depressive episode. Following US Public Health Service criteria, obese subjects were defined as those with body mass index scores at the 85th percentile or higher. Covariates were age, sex, education, marital status, social isolation and social support, chronic medical conditions, functional impairment, life events, and financial strain. Results were mixed. In cross-sectional analyses, greater odds for depression in 1994 were observed for the obese, with and without adjustment for covariates. When obesity and depression were examined prospectively, controlling for other variables, obesity in 1994 predicted depression in 1995 (odds ratio (OR) = 1.73, 95% confidence interval (CI): 1.04, 2.87). When the data were analyzed with obesity defined as a body mass index of > or = 30, cross-sectional results were the same. However, the prospective multivariate analyses were not significant (OR = 1.43, 95% CI: 0.85, 2.43). Although these data do not resolve the role of obesity as a risk factor for depression, overall the results suggest an association between obesity and depression. The authors found no support for the "jolly fat" hypothesis (obesity reduces risk of depression). However, there has been sufficient disparity of results thus far to justify continued research.