Twelve groups of eight rats and two control groups of sixteen rats were given semisynthetic diets with 40% energy as fat for a period of 76 d. All diets contained a minimum of 3% energy as linoleic acid and comparable basal levels of D-α- and D-γ-tocopherol. The diets varied in fat composition and in the content of DL-α-tocopheryl acetate. The diets high in polyunsaturated fatty acids (PUFA) were either rich in fish oil (FO; groups 1–4; 10% energy as fish oil PUFA), linseed oil (LN; groups 1–4; 10% energy as α-linolenic acid) or sunflower seed oil (SF; groups 1–4; 10 + 3% energy as linoleic acid). The control groups were given a diet high in monounsaturated fatty acids (MUFA; CO 1; 10 + 13% energy as oleic acid) or a diet with an ‘average’ linoleic acid content (CO 2; 8.5% energy as linoleic acid). Of each high PUFA diet three groups were supplemented with graded levels of DL-α-tocopheryl acetate. Steatitis, a sensitive histopathological indicator of vitamin E deficiency in animals fed on diets rich in fatty acids with three or more double bonds, was observed only in the adipose tissue of the FO groups, even in the group with the highest DL-α-tocopheryl acetate supplementation. Liver and serum α- tocopherol levels were found to be positively correlated and liver and serum γ-tocopherol levels negatively correlated with dietary DL-α-tocopheryl acetate. The groups on the FO diets had significantly reduced liver and serum tocopherol levels in comparison with the groups on the other high-PUFA diets. With the supplementation scheme used for the FO groups the liver α-tocopherol levels of both control groups were reached but the serum control levels were not.