Curcumin (diferuloylmethane), a yellow pigment that is obtained from the rhizomes of Curcuma longa Linn., is a major component of turmeric and is commonly used as a spice and food-coloring agent. The inhibitory effects of feeding commercial grade curcumin (77% curcumin, 17% demethoxycurcumin, and 3% bisdemethoxycurcumin) in AIN 76A diet on carcinogen-induced tumorigenesis in the forestomach, duodenum, and colon of mice were evaluated. Administration p.o. of commercial grade curcumin in the diet inhibited benzo(a)pyrene-induced forestomach tumorigenesis in A/J mice, N-ethyl-N'-nitro-N-nitrosoguanidine-induced duodenal tumorigenesis in C57BL/6 mice, and azoxymethane (AOM)-induced colon tumorigenesis in CF-1 mice. Dietary commercial grade curcumin was given to mice at: (a) 2 weeks before, during, and for 1 week after carcinogen administration (during the initiation period); (b) 1 week after carcinogen treatment until the end of the experiment (during the postinitiation period); or (c) during both the initiation and postinitiation periods. Feeding 0.5-2.0% commercial grade curcumin in the diet decreased the number of benzo(a)pyrene-induced forestomach tumors per mouse by 51-53% when administered during the initiation period and 47-67% when administered during the postinitiation period. Feeding 0.5-2.0% commercial grade curcumin in the diet decreased the number of N-ethyl-N'-nitro-N-nitrosoguanidine-induced duodenal tumors per mouse by 47-77% when administered during the postinitiation period. Administration of 0.5-4.0% commercial grade curcumin in the diet both during the initiation and postinitation periods decreased the number of AOM-induced colon tumors per mouse by 51-62%. Administration of 2% commercial grade curcumin in the diet inhibited the number of AOM-induced colon tumors per mouse by 66% when fed during the initiation period and 25% when fed during the postinitiation period. The ability of commercial grade curcumin to inhibit AOM-induced colon tumorigenesis is comparable to that of pure curcumin (purity greater than 98%). Administration of pure or commercial grade curcumin in the diet to AOM-treated mice resulted in development of colon tumors which were generally smaller in number and size as compared to the control group of AOM-treated mice. These results indicate that not only did curcumin inhibit the number of tumors per mouse and the percentage of mice with tumors but it also reduced tumor size. Histopathological examination of the tumors showed that dietary curcumin inhibited the number of papillomas and squamous cell carcinomas of the forestomach as well as the number of adenomas and adenocarcinomas of the duodenum and colon.