A retrospective cohort mortality study was conducted among 8147 men and 627 women employed in a gray iron foundry for at least 6 months between 1950 and 1979. More than 1700 deaths occurred during a 35-year period of observation. Standardized mortality ratios (SMRs) for all causes were close to expected values based on the US general population as the standard. The mortality of nonwhite men was significantly increased for lung cancer (SMR 132) and ischemic heart disease (SMR 126). Other moderate, but nonsignificant excesses were noted among nonwhite men for cancers of the stomach, pancreas, and prostate, for diabetes mellitus and pulmonary emphysema, and among white men for cancers of the lung and stomach, gastric and duodenal ulcers, pulmonary emphysema, and suicide. Small mortality increases were observed in both racial groups for cerebrovascular disease. The lack of a trend with time since hire and duration of foundry employment suggests that lung cancer mortality may not be associated with exposure to the foundry environment. Utilizing indirect measures of smoking, it appears that virtually all excess lung cancer deaths among whites and at least some of the excess among nonwhites could be explained by smoking habits. Similarly, smoking may have been responsible for the mortality excesses from emphysema, cerebrovascular diseases, and ischemic heart disease.