The authors studied 2,950 population-based colon cancer cases in males in Los Angeles County, California, that were diagnosed between 1972 and 1981. To determine if colon cancer risk is reduced by physical activity on the job in males aged 20-64 years, the authors first rated each occupation by judging the activity level as high, moderate, or sedentary. Men with sedentary jobs had a colon cancer risk at least 1.6 times that of men whose jobs required a high level of activity. Risk increased in a stepwise manner as activity level decreased. This gradient was consistently seen within each socioeconomic stratum, among whites, blacks, immigrant and native Hispanics, and for each subsection of the colon from the hepatic flexure to the sigmoid. The protective effect of physical activity was very strong in the descending colon and diminished in a gradient both proximally and distally. There was no such relationship between physical activity and risk for rectal cancer. Physical activity may play a major, previously unrecognized role in colon cancer etiology. Such a role is consistent with the known pattern of colon cancer occurrence and with our understanding of colon physiology and colon cancer pathogenesis. In addition to the implications for prevention, understanding the effects of physical activity on colon cancer risk may allow future studies to evaluate more accurately the role played by diet.