We investigated the association between self-reported alcohol ingestion and colorectal cancer in a cohort of male smokers in Finland. Among 27,109 men aged 50 to 69 years, 87 colon and 53 rectal cases were diagnosed during the five to eight years of follow-up. Among drinkers, colorectal cancer risk increased with the amount of alcohol consumed (P trend = 0.01) with risk increasing by 17 percent for each drink consumed. Both beer and spirits contributed to this increased risk. Further analyses revealed that the positive association with alcohol was primarily for colon cancer (P trend = 0.01). Interestingly, risk of colorectal cancer associated with drinking (cf self-reported abstinence) changed with follow-up time, suggesting an inverse association for alcohol early in follow-up, and a positive association after about three-and-a-half years of follow-up. Follow-up time did not modify the positive association with amount of alcohol among drinkers, however. Results also indicated that beta-carotene supplementation may attenuate the effect of alcohol on colorectal cancer risk among drinkers. In conclusion, this study supports a role for alcohol in colon carcinogenesis and suggests that similar studies should evaluate carefully the effects of lifetime drinking habits and recent abstinence.