Duplicate groups of rainbow trout (Salmo gairdneri) (mean weight 11 g) were given for 40 weeks one of four partially purified diets that were either adequate or low in selenium or vitamin E or both. Weight gains of trout given the dually deficient diet were significantly lower than those of trout given a complete diet or a diet deficient in Se. No mortalities occurred and the only pathology seen was exudative diathesis in the dually deficient trout. There was significant interaction between the two nutrients both with respect to packed cell volume and to malondialdehyde formation in the in vitro NADPH-dependent microsomal lipid peroxidation system. Tissue levels of vitamin E and Se decreased to very low levels in trout given diets lacking these nutrients. For plasma there was a significant effect of dietary vitamin E on Se concentration. Glutathione (GSH) peroxidase (EC 1.11.1.9) activity in liver and plasma was significantly lower in trout receiving low dietary Se but was independent of vitamin E intake. The ratios of hepatic GSH peroxidase activity measured with cumene hydroperoxide and hydrogen peroxide were the same for all treatments. This confirms the absence of a Se-independent GSH peroxidase activity in trout liver. Se deficiency did not lead to any compensatory increase in hepatic GSH transferase (EC 2.5.1.18) activity; values were essentially the same in all treatments. Plasma pyruvate kinase (EC 2.7.1.40) activity increased significantly in the trout deficient in both nutrients. This was thought to be due to leakage of the enzyme from the muscle and may be indicative of incipient (subclinical) muscle damage.