An international comparison study of the relationship among dietary predictors, tobacco consumption, income, and truncated age-adjusted lung cancer mortality was conducted with the use of time-lagged data available for 43 countries. A regression analysis weighting each country by the square root of the appropriate population and adjusting for several covariates showed that calories from dietary fat were highly significantly associated (P less than .0001) with lung cancer mortality. This finding was obtained after accounting for disappearance data for tobacco (P less than .0001), the dominant risk factor for lung cancer, and total nonfat calories (P less than .002). Other covariates included per capita gross national product and proportion of calories from fruits, vegetables, roots, tubers, and pulses. Weaknesses of this study, statistical considerations, biologic plausibility, in particular the modulation of tobacco-smoke carcinogens by lipids in the lung, and suggestions for future study are discussed.