A previous report suggested that the cancer risk associated with hard liquor is greater among users of dark liquors, which may contain more potentially carcinogenic compounds than light liquors. To test this hypothesis, we examined alcohol consumption data obtained by interview with 921 cases and 900 controls who participated in 1984-1985 in a population-based case-control study of oral and pharyngeal cancer conducted in 4 areas of the United States. Among heavy liquor drinkers (30+ drinks/week), odds ratios were 13.2 (95% CI = 5.2-33.5) for those who usually drank light liquors vs. 4.6 (95% CI = 2.7-7.9) for those who usually drank dark liquors, with higher risks for light vs. dark liquors at each anatomical subsite. This analysis provides no support for the notion that dark liquors are more carcinogenic than light liquors or that non-ethanolic ingredients of alcoholic beverages are major contributors to the excess risk of oral and pharyngeal cancer.