The relationship of alcohol use to diet was examined in 2272 male and 2337 female adults aged 45 y and older who provided a quantitative diet history during 1977-1979. Mean values for each dietary variable, adjusted for smoking, ethnicity, income, and education, were compared in each sex between abstainers and drinkers and by tertile of ethanol intake. Linear relationships with extent of drinking were also sought. Drinkers were found to be less obese than abstainers. Consumption of carbohydrate, vitamins, calcium, fruits, fruit juices, and raw vegetables was greater among abstainers whereas consumption of fat (particularly polyunsaturated fatty acids), cholesterol, zinc, meat, pickled vegetables, and dried fish was greater among drinkers. Because validations of dietary questionnaires have shown that alcohol consumption is more accurately recalled than food intake, the possibility of a residual confounding effect for these dietary variables should be considered in epidemiologic studies of alcohol and health.