The effect on oral cancer risk of different types of alcoholic beverage was investigated using data from a hospital-based case-control study. Owing to the small numbers of subjects drinking one beverage exclusively, it was necessary to classify drinkers as consumers of predominantly beer, wine, or hard liquor (i.e., more than 50% of their whiskey equivalents of alcohol derived from a specific beverage). The number of predominantly wine drinkers was too small to permit analysis. Logistic regression was used to obtain estimates of the risk associated with each predominant beverage, with adjustment for other risk factors and confounding variables. In males, the odds ratio for predominantly beer drinkers increased with increasing level of intake, reaching 4.87 (95% confidence interval: 2.51-9.46) in drinkers of 7+ oz. of whiskey equivalents/day. The odds ratio for predominantly hard liquor drinkers showed a similar increase, reaching 5.74 (95% confidence interval: 2.94-11.22) in predominantly hard liquor drinkers consuming 7+ oz. of whiskey equivalents/day, suggesting that the effect of these 2 major types of alcoholic beverage is of similar magnitude. The trends were less clearcut in women due to small numbers of drinkers.