Life-long bowel habits of 685 colorectal cancer cases and 723 age/sex frequency matched community controls were investigated as one part of a large, comprehensive, population-based study of colorectal cancer incidence, etiology, and survival, The Melbourne Colorectal Cancer Study. Self-reported chronic constipation was statistically significantly more common in cases than in controls (P = .05). Three or more bowel actions per day were reported by more cases than controls but the total number of respondents in this subset consisted of only ten cases and two controls. Otherwise, the frequency and consistency of bowel motions was similarly distributed among cases and controls. Constipation disappeared as a significant risk when simultaneously adjusted for previously determined dietary risk factors, indicating that it is the diet and not the constipation that is associated with the risk of large-bowel cancer. Additionally, a highly statistically significant association (P = .02) was found with the risk of colorectal cancer in those who reported constipation and also had a high fat intake, a finding consistent with current hypotheses of colorectal carcinogenesis. It is concluded that chronic constipation, diarrhea, and the frequency and consistency of bowel motions, as well as laxative use, are unlikely to be etiologic factors in the development of colorectal cancer. Self-reported chronic constipation is a marginally significant indicator of excess risk of large-bowel cancer and may be used as one of the indices in the screening of individuals for this cancer.