Seven groups of female Sprague-Dawley rats (˜200 gm initial body weight) were injected i.p. with a single subcarcinogenic dose of diethylnitrosamine (40 mg per kg body weight) between 8 to 10 hr after partial hepatectomy, and after a recovery period of 3 weeks (herein called induction stage) received 0.05% phenobarbital in the diet for the rest of the experiment (promotion stage). The rats were fed a 20% casein-based diet containing 0.16 ppm of selenium or the same diet supplemented with 4 or 6 ppm of selenium as sodium selenite. The effects of these three dietary regimens were tested when administered 9 to 11 days before and during induction, 1 week before and during promotion or during the entire experiment. Pair-feeding conditions were used to minimize influences due to differences in food intake and growth. Despite similarities in food intakes, the growth rates in groups receiving the 6 ppm-selenium diet during promotion or during the entire experiment were in general significantly lower than in rats fed the 4 ppm-selenium diet or the 0.16 ppm-selenium basal diet. Survival rates were also significantly reduced in rats fed the 4 and 6 ppm-selenium diets during promotion or during the entire experiment. In rats killed at the 19th week for interim assessment of the experiment's progress, the stereologically analyzed numerical and volumetric densities of hepatic premalignant hyperplastic nodules did not differ significantly between groups. All the remaining rats were killed at the 46th week. The statistical analysis of the total incidence of hepatic carcinomas and renal tumors encountered in rats during the course of the experiment and in those killed at the end of the study showed no significant differences between groups. Toxic effects of selenium on liver structure were quite conspicuous in rats under prolonged treatment with 4 and 6 ppm-selenium diets.