Cause specific mortality was surveyed among 37,682 male employees with three or more days of service between 1940 and 1982 at the Midland or Bay City, Michigan, locations of Dow Chemical USA. Vital status was ascertained through 1982 for 97.5% of the cohort members, and death certificates were obtained for 97.1% of the 7,751 decedents. Comparisons of observed mortality with expected levels based on any of three general population groups (US, Michigan, or seven local counties) consistently demonstrated lower mortality in the cohort from each of the major causes of death, including total malignant neoplasms. Unique among hourly employees was significant excess mortality in the categories of cancer of other lymphatic tissue, and motor vehicle accidents, and both hourly and salaried nonexempt employees experienced significantly higher mortality from other and ill-defined cancers. The influence of duration of employment and age at and period of hire were explored with the Mantel-Haenszel method as adapted for a cohort study. Results were evaluated both including and excluding the mortality experience of subsets of employees with past exposure to known human carcinogens (arsenic, asbestos, bis-chloromethyl ether, benzene, organic dyes, and vinyl chloride). The use of the general mortality survey in monitoring whether or not there are major health problems among the employees and in setting research priorities is emphasized.