We examined the relation between self-reported physical activity and large bowel cancer in a prospective cohort of men and women who participated in the Framingham Study. Self-assessments of physical activity were available from the fourth biennial examination on a total of 1906 men and 2308 women aged 30 to 62 yr in 1954. The cohort was followed for up to 28 yr and yielded 152 cases (73 men, 79 women) of large bowel cancer. Inactivity was associated with an increased risk of large bowel cancer among men but not among women. The relative risk estimates for large bowel cancer among men in the middle and lowest tertiles of a physical activity index (compared with the highest tertile) were 1.4 (95% confidence intervals, 0.8-2.6) and 1.8 (1.0-3.2), respectively. Among women the comparable estimates were 1.2 (0.7-2.1) and 1.1 (0.6-1.8), respectively. These findings were unchanged after adjustment for body mass index, serum cholesterol, alcohol, and other potentially confounding variables. The narrow range of physical activity and the minimal heavy activity reported by women in this cohort may have limited our ability to detect an association between physical activity and large bowel cancer among women.