The Center for Epidemiologic Studies Depression scale (CES-D) has been widely used in studies of late-life depression. Psychometric properties are generally favourable, but data on the criterion validity of the CES-D in elderly community-based samples are lacking. In a sample of older (55–85 years) inhabitants of the Netherlands, 487 subjects were selected to study criterion validity of the CES-D. Using the 1-month prevalence of major depression derived from the Diagnostic Interview Schedule (DIS) as criterion, the weighted sensitivity of the CES-D was 100%; specificity 88%; and positive predictive value 13·2%. False positives were not more likely among elderly with physical illness, cognitive decline or anxiety. We conclude that the criterion validity of the CES-D for major depression was very satisfactory in this sample of older adults.