A nested case‐control study of lung cancer was conducted among workers at an iron foundry and two engine manufacturing plants whose lung cancer mortality rates were slightly higher than expected. The study included 231 lung cancer cases and 408 controls for whom complete work histories were obtained. There was no association between usual plant of employment and lung cancer mortality. The odds ratio for persons employed for 20 or more years in the foundry compared with persons employed in the engine plants was 0.90 (95% confidence interval: 0.55, 1.5). Long‐term employment as an engine plant worker was associated with odds ratios slightly, but not statistically significantly, below unity. In the foundry, only usual employment in the material handling departmental group and any employment in the quality control departmental group were statistically significantly directly related to lung cancer risk. However, the number of subjects so employed was small and there was no dose‐response relation between length of employment in these departmental groups and lung cancer risk. Cases were less frequently employed than were controls in engine plant machining and assembly jobs and departments. It is concluded that employment in this facility was either unrelated, or only weakly related, to lung cancer risk. Am. J. Ind. Med. 31:414–421, 1997. © 1997 Wiley‐Liss, Inc.