Mortality at two engine plants was analyzed using proportional mortality and logistic regression models of mortality odds ratios to expand previous observations of increased cancers of the stomach, pancreas, and bladder, and cirrhosis of the liver among workers exposed to machining fluids. Causes of death and work histories were available for 1,870 decendents. There was a significant excess of deaths coded as diabetes for white men in both plants (PMR = 25/16.7 = 1.5, 95% CI = 1.02, 2.20), and a deficit of respiratory diseases. Black men had fewer than expected diabetes deaths and more emphysema deaths. Elevated PMRs for cancers of the stomach, pancreas, prostate, bladder, and kidney were not statistically significant in plantwide populations. However, stomach cancer mortality increased with duration in camshaft and crankshaft production at Plant 1 (OR = 5.1, 95% CI = 1.6, 17; at mean duration of exposed cases), and among tool room workers (OR = 6.3, 95% CI = 1.3, 31), but these results were based on five cases. Nitrosamines were probably present in camshaft and crankshaft grinding at Plant 1. Pancreas cancer risk increased among workers at both plants ever employed in inspection (OR = 2.5, 16), in machining with straight oil (OR = 3.6, 95% CI = 1.04, 12), or in skilled trades (OR = 2.9, 95% CI = 1.1, 7.5). Lung cancer increased in cylinder head machining (OR = 3.9, 95% CI = 1.4, 11), millwright work (OR = 3.8, 95% CI = 1.6, 9.0), and in Plant 2 generally (OR = 1.45, 95% CI = 0.97, 2.2). Potential lung carcinogens included heat treatment emissions, chlorinated oils, and coal tar fumes (millwrights). Bladder cancer increased with duration among workers grinding in straight oil MF (OR = 3.0, 95% CI = 1.15, 7.8) and in machining/heat-treat operations (OR = 2.9, 95% CI = 1.14, 7.2).