Male mice were fed a torula yeast-based diet containing different amounts of added selenium for a period of 4 months. Liver glutathione peroxidase activity assayed with H 2O 2 showed a logarithmic dependence on dietary selenium with a saturation plateau above 2 ppm Se and an extrapolated zero of 0.02 ppm Se. In contrast, liver selenium content and GSH-Peroxidase activity showed a linear correlation. Glutathione peroxidase activity became undetectable at a liver Se content of about 90 ng Se/g liver wet wt. Thus, about 10% of liver selenium is not related to GSH-Px activity. Five dietary groups were supplemented, respectively, with 0, 0.05, 0.5, 5.0 and 10 ppm Se in the form of Na 2SeO 3. Some changes in drug metabolism enzymes were observed with the high Se diets. An increase occurred in Non-Se-GSH activity as well as in ethacrynic acid-assayed GSH transferase, these are interpreted as early signs of Se toxicity. The diet containing 0.01 ppm Se with no supplementary Se produced the multiple hepatic enzyme modulations which were previously reported. The animals raised on this very low Se diet had normal hepatic contents of glutathione, α-tocopherol, calcium, magnesium, iron, zinc, copper and manganese compared to controls supplemented with 0.5 ppm Se. However, significant changes in the microsomal fatty acid pattern were observed while the total phospholipid content as well as membrane fluidity showed no differences between the two dietary groups.