A 120-day assay was made of Se concentration and glutathione peroxidase Se-dependent (GSHpx) activity in plasma, erythrocytes and liver of female Sprague-Dawley rats fed either a Torula diet (low Se content) or a durum wheat diet providing 28–405 µg of Se/kg diet. For all groups there was a strong increase for the first 20 days in plasma and liver Se; for the remaining period the increase was lower; erythrocyte Se increased, as before, in the first 60 days for groups fed 28–200 µg/kg diet, after this period it decreased in the groups fed high-Se diet. Plasma GSHpx activity was extremely sensitive to dietary Se levels and increased for up to 40 days of repletion, after which the trend was to plateau; liver and erythrocyte GSHpx activity increased continuously for up to 60 days of supplementation, it then remained constant. The correlation between Se in plasma, liver and erythrocytes and its GSHpx activity was statistically significant (p < 0.001). These results suggested that in long-term-fed rats, a diet with levels of 100–400 µg/kg was sufficient to satisfy the rat’s needs; there was no evidence of toxicity and, moreover, in all tissues, an amount of Se, probably bound to proteins with unknown functions, w