To determine the prevalence of cardiac disorders as risk factors for stroke, we conducted a survey in 1986 in a stratified random sample of the population of Rochester, Minnesota, 35 years of age or older. The medical records of the 2,122 subjects in the sample were retrieved with use of the Rochester Epidemiology Project medical records linkage system. The data were used to estimate (1) the reliability of self-reported information about cardiac and cerebrovascular disorders and (2) the age- and sex-specific prevalence of diabetes mellitus and various cardiac and cerebrovascular conditions. The estimated prevalence for selected risk factors in the population 35 years of age or older was 5.8% for diabetes mellitus, 3.3% for myocardial infarction, 1.2% for mitral valve disease, 4.2% for left ventricular hypertrophy, and 2.8% for atrial fibrillation or flutter. These data can be used to estimate resources required for evaluation and management of the disorders. When the prevalence and the relative risk for stroke are known for a particular cardiac disorder, the proportion of stroke attributable to that disorder can be estimated.