Coffee has been observed to be associated weakly or not at all with bladder cancer risk, inversely with colon cancer risk, and inconsistently with rectal cancer risk. The association between these cancers and consumption of coffee and tea was examined in a single case-control study conducted in Ontario, Canada from 1992 to 1994. A questionnaire was filled out by 927 bladder cancer cases, 991 colon cancer cases, 875 rectal cancer cases, and 2118 population controls. Although bladder cancer risk was not associated with coffee or tea, risk estimates associated with coffee among subjects who had never smoked were non-significantly increased. Colon cancer risk was inversely associated with coffee. Relative to those drinking less than 1 cup of coffee per day, the odds ratios (OR) for those drinking 1-2 cups was 0.9 (95% CI 0.8-1.1), for those drinking 3-4 cups was 0.8 (95% CI 0.7-1.0), and for those drinking 5 or more cups was 0.7 (95% CI 0.5-0.9); these ORs decreased linearly (P = 0.008). The reduced risk estimates were more pronounced with cancer of the proximal colon than the distal colon. Rectal cancer risk was not associated with either coffee or tea. Coffee consumption was observed to have a different relationship for each of the cancer sites and tea consumption was not related to any cancer site.