OBJECTIVE: To estimate the effects of alcohol consumption on the risk of colorectal cancer according to anatomical subsite. METHODS: Between 1979 and 1985 a population-based case-control study of cancer at multiple sites was carried out in Montreal. This analysis was restricted to the 585 cases with adenocarcinoma of the large bowel, aged 35-70 years, who underwent face-to-face interviews. Controls (n = 500) were selected either from electoral lists or by random-digit dialing and were frequency-matched to the cases on age. Polytomous logistic regression was used to estimate the risk of cancer of the proximal colon, distal colon, and rectum in relation to the consumption of alcoholic beverages. RESULTS: Daily consumption of alcohol of any type was associated with increased risks of cancer of the distal colon [odds ratio (OR)= 2.3; 95% confidence interval (CI) 1.4-3.7] and the rectum (OR = 1.6; 95% Cl 1.0-2.6), but not with an increased risk of cancer of the proximal colon (OR = 1.0; 95% CI: 0.6-1.7). When type of beverage was considered, beer showed strong associations with cancer at all three subsites (ORs among the heaviest drinkers ranging from 1.8 to 2.4), spirits showed weaker associations with cancer at all three subsites (ORs ranging from 1.4 to 1.6), and wine showed null associations. CONCLUSIONS: The results are consistent with the hypothesis that consumption of alcoholic beverages increases the risk of colorectal cancer. The evidence is strongest for effects on the distal colon and rectum, and, among the three types of beverage, it most strongly implicates beer.