AbstractFour groups of weanling male Wistar rats (Groups A—D) received diethylnitrosamine (DEN, 40 ppm) in their drinking water for four weeks; after a recovery period of two weeks, they received (for the rest of the experiment) phenobarbital (PB, 500 ppm) added to a Torula yeast‐based diet containing 0.17 ppm of selenium. Dietary selenium (2 ppm), as sodium selenite, was given to Group B one week before and during DEN treatment, to Group C one week before and during PB treatment, and to Group D during the entire experiment. Groups A and E received the unsupplemented diet, whereas Group E was not treated with DEN or PB. Pair‐feeding conditions were used to minimize possible influences of differences in food intake and growth. Rats were killed at the 19th and 24th weeks after the experiment began. No significant differences were found in food and fluid intakes or in growth rates among the groups. Livers in Group E were histologically normal, whereas preneoplastic and neoplastic lesions were found in all other groups. In rats killed at the 19th and 24th weeks, the numerical and the volumetric densities of preneoplastic lesions did not differ significantly between all the groups. Similarly, the incidence of hepatocellular carcinomas only detected at 24 weeks was not significantly different between the groups. These results indicated that in this particular model of hepatocarcinogenesis, the dietary supplementation of 2 ppm of selenium did not modify the development of preneoplasia and carcinomas.