To investigate lung cancer risk, the authors conducted a historical cohort mortality study of 4,459 mild steel welders who had been employed at three midwestern plants which manufactured heavy equipment. Follow-up began in the mid-1950s and extended through 1988. All welders had at least 2 years welding experience (average duration, 8.5 years). This cohort had no occupational exposure to asbestos or stainless steel fumes (containing nickel and chromium), two potential confounders in previous welders studies. A comparison population of 4,286 nonwelders, all with at least 2 years employment at the same plants, was also studied. Nonwelders had never been welders and were allowed to have no more than 90 days employment as a painter, foundryman, or machinist. Sampling data collected from 1974-1987 indicated that welders were exposed to 6-7 mg/m3 of total particulate and 3-4 mg/m3 of iron oxide, while nonwelders had negligible exposures to welding fumes. When compared with the United States population, both welders and nonwelders had elevated rates for lung cancer (standardized mortality ratios (SMRs): welders, SMR = 1.07; nonwelders, SMR = 1.17), but neither SMR was significantly elevated. Limited smoking data based on a 1985 survey indicated that both welders and nonwelders smoked more than the United States population, possibly accounting for part of their elevated lung cancer rates. There was no trend of increased risk for welders with increased duration of exposure. The only other cause of death significantly elevated was emphysema among welders. Nonmalignant respiratory disease was not elevated for welders (SMR = 0.96). When welders were compared with nonwelders directly for lung cancer, the rate ratio was 0.90.