This study was undertaken to estimate the point-prevalence of Research Diagnostic Criteria (RDC) depressive syndromes in Alzheimer's disease (AD) and to evaluate the validity of existing and potential alternative diagnostic criteria for major depression in the presence of probable AD. Twenty-six subjects with probable AD of mild to moderate severity and their caregivers were interviewed to estimate the prevalence of RDC depressive syndromes. For the evaluation of the validity of RDC for major depression, an additional 8 probable-AD subjects with suspected depression were added to the sample. Sensitivity, specificity, and correlation with diagnosis of RDC major depression were calculated for each diagnostic criterion, and existing major depressive criteria were compared to potential alternative criteria currently used for RDC minor depression. Of the subjects in our prevalence sample, 15.4% were found to have major depression; 23.1%, minor depression; and 11.5%, intermittent depression. In our validation sample, two criteria for major depression, self-reproach/guilt and thinking/concentration difficulty, were weakly associated with the final diagnosis of major depression because of poor sensitivity or specificity. In contrast, three possible alternative criteria were significantly associated with the diagnosis of major depression and showed high sensitivity and specificity. These included nonverbal manifestations of depression, irritability/complaining, and demandingness/dependency. We conclude that RDC depressive syndromes are common in probable AD of mild to moderate severity. In the presence of AD, the validity of some existing major depressive criteria may be limited in comparison to several potential alternative criteria because of relatively poor sensitivity and/or specificity.