In a prospective study of a sample of male alcoholics, age standardized rates of death from cancer and other causes were compared with expectancies based on the mortality of the general male population of Ontario and that of U. S. veterans in the Dorn Study. A typical profile of mortality due to alcoholism was found with high excess mortality from cirrhosis, pneumonia, violent causes, lung cancer, and cancers of the upper digestive and respiratory tracts. There was no evidence of the associations recently reported in the literature between alcohol use and other cancers such as those of the stomach, colon, and pancreas. Comparison with veterans whose smoking resembled that of the alcoholics revealed similar rates of death from lung cancer, considerable excess mortality among the alcoholics from cancers of the upper digestive and respiratory tracts, and no difference in overall cancer mortality. Heavy alcohol use per se increases the risk of cancer at certain sites, but it may not increase the overall risk of neoplastic disease.