The blood selenium (Se) concentration in the U.K. population has declined by approx. 50% between 1974 and 1991, reflecting a large decrease in dietary Se supply, with intakes only half the reference nutrient intake of 1 µ g/kg body weight. Tissue levels of Se are readily influenced by dietary intake. Therefore selenoprotein activity may be sub-optimal due to low Se status, and thus compromise normal cell function. To examine the effects of changing Se intake on selenoproteins, we have determined the relative effectiveness of organic selenomethionine and inorganic sodium selenite (50 µ g of Se daily for 28 days) in modulating glutathione peroxidase activities in blood cells from 45 healthy men and women, from a U.K. population. Transient and acute changes in lymphocyte, granulocyte and platelet phospholipid-hydroperoxide glutathione peroxidase (GP×4) activity occurred by day 7 or 14 of sodium selenite treatment and by day 7 in lymphocytes from selenomethionine-treated subjects compared with controls taking a placebo. In contrast, GP×4 activity in granulocytes and platelets in the selenomethionine group increased gradually over the 28 days. Cytosolic glutathione peroxidase (GP×1) activity in these blood cells from both treatment groups increased gradually over the 28 days. For each cellular selenoenzyme activity a significant inter-individual difference ( P < 0.001) in the extent of the response to Se supplementation was observed, but this was not related to blood Se concentrations either before or after treatments. Significant inverse correlations were evident between baseline enzyme activities and percentage change in activity after 28 days of supplementation [e.g. lymphocyte GP×4, r =-0.695 ( P < 0.001)], indicating that pre-treatment activity may be sub-optimal as a result of poor Se status. The different and contrasting effects that Se supplementation had on blood selenoenzyme activities may be indicative of a difference in metabolic need for Se regulated at the level of Se-dependent cell function.