Diet is a major influence on the responses of experimental animals to drugs, toxins, and carcinogens. Two diets used widely in toxicological and carcinogenic studies, and considered to be nutritionally adequate, were compared with respect to neoplastic responses to 2-acetylaminofluorene (2-AAF). Both sexes of weanling B6C3F1 mice were fed either AIN-76A (a purified diet) or NIH-07 (a natural ingredient diet), with or without 2-AAF, for up to 2 years. Dosages of 2-AAF were administered to males at 0, 40, 60, or 80 ppm in each diet and to females at 0, 150, 200, or 250 ppm. Each group consisted of 96 mice, except the groups of females dosed at 0 and 150 ppm, which consisted of 120 and 72 mice, respectively. The incidence of malignant liver tumors was significantly greater in all AIN-fed groups compared to corresponding NIH-fed groups, as was the total incidence of tumors (malignant + benign). Similarly, the incidence of malignant and total bladder tumors was greater in AIN-fed groups of mice administered the two high doses of 2-AAF for each sex compared to NIH-fed groups. No malignant bladder tumors were observed among any of the groups of mice receiving control diets. This was also true for the males on low dosages (40 ppm) of 2-AAF. The AIN-fed group of low dose females (150 ppm) developed 4% incidence of malignant bladder tumors. This study dramatically showed the importance of diet selection in chronic carcinogenic studies and suggests that results obtained with the use of a purified diet may differ both qualitatively and quantitatively from results obtained with a natural ingredient diet.