The alcohol intake of a cohort of Japanese men in Hawaii is directly and significantly related to the risk of developing rectal cancer, whether assessed on the basis of amount consumed or as a percent of total calories. Wine and whiskey are directly related to rectal cancer, but beer is the only alcoholic beverage that displays a statistically significant dose-response (P = 0.008). Colon cancer risk also is related directly to alcohol intake, but the association is statistically significant only when measured as a percent of energy intake. This suggests that alcohol might displace cancer inhibitors from the diet. Calcium, vitamin C, and dietary fiber are inversely related to colon cancer risk in this cohort, and each of these micronutrients displays statistically significant negative correlation with alcohol intake. A possible positive association between alcohol and lung cancer was ruled out after adjusting for cigarette smoking. Cancers of the prostate and stomach were unrelated to alcohol intake, but the risk of acquiring cancer at all other sites combined was strongly related to alcohol intake.