• Using longitudinal data from a community study of 9900 adults drawn from four sites in the United States and interviewed twice, 1 year apart, we investigated the predictors of first-onset major depression. Using odds ratios to estimate relative risk, we found that persons with depressive symptoms, compared with those without such symptoms, were 4.4 times more likely and persons with dysthymia were 5.5 times more likely to develop a first-onset major depression during a 1-year period. The lifetime prevalence rate for depressive symptoms was 24%. The attributable risk is a compound epidemiologic measure that reflects both the relative risk associated with depressive symptoms (4.4) and the prevalence of exposure to that risk (24%). It is a useful measure to document the burden of a risk to the community, and it was determined to be greater than 50%. Thus, more than 50% of cases of first-onset major depression are associated with prior depressive symptoms. The high prevalence of depressive symptoms in the community and their strong association with first-onset major depression make them important from a public health perspective. Because depressive symptoms are often unrecognized and untreated in clinical practice, we conclude that their identification and the development of effective treatments could have implications for the prevention of major depression.