The reported beverage habits of 569 lung cancer patients and 569 control patients admitted to Roswell Park Memorial Institute (RPMI) were studied, 355 male cases and 214 female cases being matched to controls within strata of age and residence. Smoking history and an index of vitamin A from vegetables had significant, dose-response associations with risk. Animal fat intake as measured by an index of animal fats from meats showed elevated risks which were not significant. Three vegetables rich in vitamin A and 3 meats contributing to the animal fat index were, individually, associated with lung cancer risk. Frequency of consumption of milk, coffee, tea, soft drinks and alcoholic beverages was studied in multiple logistic regression analyses which controlled for smoking history, intake of vitamin A from vegetables and education level. Subjects reporting consumption of whole milk 3 or more times daily had a 2-fold increase in lung cancer risk compared to those who reported never drinking whole milk (RR = 2.14). The same frequency of intake of reduced-fat milk was associated with a significant protective effect (RR = .54). Significant risk variations were observed for other beverages but, with the exception of frequencies of reported diet cola and decaffeinated coffee intake, dose-response patterns were not evident.