The association between occupational physical activity and risk of colorectal cancer by age and anatomic site was investigated in a study of 2,503 males with colorectal cancer registered with the New Zealand Cancer Registry during 1972-80. Occupational groups that involved high levels of physical activity or were predominantly sedentary were identified prior to analysis of the registry data. Relative to males in high physical activity occupations, males in sedentary occupations had an increased incidence of both cancer of the colon (relative risk [RR] = 1.2; 95 percent confidence interval [CI] = 1.0-1.4) and rectum (RR = 1.3, CI = 1.0-1.5). The RRs for sedentary workers were particularly elevated in the 35-44 and 45-54 year age-groups for colon cancer (RR = 1.8 and 1.5, respectively) and in the 45-54 year age-group for rectal cancer (RR = 1.5), whereas there was no increase in risk for sedentary workers in the 55-64 year age-group for either cancer site. The general increase in colon cancer incidence for New Zealand during the study period was reflected in the sedentary group, but there was no change in incidence among men in occupations involving high or intermediate levels of physical activity. There was no obvious pattern for the increased cancer risk for men in sedentary occupations by anatomic site. Current physiologic hypotheses for the effect of physical activity on colon cancer risk do not adequately explain an association of physical activity with risk of rectal cancer.