Using the Cancer-Environment Registry of Sweden, which links census information (1960) with cancer incidence data (1961 to 1979), we conducted a systematic, population-based assessment of colon cancer incidence among cohorts defined by industry and occupation for all employed persons in Sweden. Small but statistically significant excesses of colon cancer were observed among white-collar occupations, including administrators, professionals, and clerical and sales workers, whereas a reduction in incidence was found among workers in agricultural and related jobs, such as farmers, fishermen, and hunters. Analysis by subsite within the colon revealed little difference in results. The observed risk patterns are consistent with previous reports on colon cancer risk and occupational physical activity levels, ie, elevated risk among sedentary white-collar workers and reduced risk among agricultural workers. Few craftsman and production processing jobs were linked to colon cancer, although statistically significant excesses were observed among shoe and leather workers, metal smiths, and foundry workers in the metal manufacturing industry. The findings indicate that occupation in general is likely to play a relatively small role in colon cancer etiology, with perhaps its major contribution an indirect one via physical activity.