Information about occupation and industry of cancer patients is routinely recorded in the Cancer Registry of Tianjin, People's Republic of China. Occupational risk factors of lung cancer for the period of 1981-1987 were assessed using other cancers as controls. For each occupation and industry, workers employed in it were considered exposed and were compared with all other workers. Elevated risks were found for a number of industries for both sexes, including, in particular, the textile, wood, metal, and construction industries. Results by occupation were similar to those by industry. Smoking scores were calculated for various occupations on the basis of a separate population survey in Tianjin. Smoking cannot explain the elevated odds ratios in males, while among females a strong correlation was observed between smoking scores and lung cancer odds ratios, suggesting a confounding effect.