Mortality among workers with 2 or more years employment at an automotive stamping and assembly complex was analyzed using standardized mortality ratio (SMR), proportional mortality ratio (PMR), and mortality odds ratio (MOR) methods. The stamping plant all-cause SMR was considerably less than expected (for white men, SMR = 0.65, 95% confidence interval [CI] = 0.54, 0.79; for black men, SMR = 0.73, 95% CI = 0.45, 1.13), indicating a strong "healthy worker effect." However, six stomach cancer deaths produced an SMR of 4.4 (95% CI = 1.62, 9.6) and a PMR of 6.8 (95% CI = 2.5, 15). Based on small numbers of cases, stomach cancer risk increased with duration in stamping and tool and die departments where exposures included drawing compound and other metalworking fluids. Stamping plant lung cancer mortality was elevated among production welders (MOR = 2.7, 95% CI = 1.2, 6.3), and increased with duration. Welding was performed on sheet metal sometimes coated with drawing compound, primer, or epoxy resin adhesive. As was observed for the stamping plant, the all-cause SMR for the two assembly plants was unusually low (for white men, SMR = 0.64, 95% CI = 0.56, 0.73; for black men, SMR = 0.57, 95% CI = 0.43, 0.75). The lung cancer SMR was not elevated but the MOR was (MOR = 1.58, 95% CI = 1.1, 2.4) and increased with assembly plant duration (MOR = 1.78, 95% CI = 1.02, 3.1, at mean duration of cases). In the assembly plants, paint oven stack emissions had been reintroduced into the plant by the ventilation system.