Weanling diabetes-prone BB rats were fed AIN-76 diets containing high (HE, 1 g/kg diet), basal (NE, 0.2 g/kg) or low (LE, trace) vitamin E and were killed at 21, 42 or 60 days of age. Plasma and tissues (adrenals, pancreas, spleen, thymus, liver, brown and white adipose tissue, muscle and testes) were analysed for vitamin E. Vitamin E levels reflected the level in the diet and no diabetic animals were detected at these times. In a second experiment, a total of 90 diabetes-prone BB rats were kept on diets LE and HE for 6 months or until they became diabetic. 11/45 on LE and 5/45 on HE became diabetic. Again, plasma and tissue levels of vitamin E reflected the levels in the diet with the exception of the thymus of diabetic rats fed the high vitamin E diet. Thymus vitamin E levels (microgram/g tissue) were 1.8 and 1.2 in LE-fed diabetics and asymptomatic rats, respectively; and 22.7 and 49.5 in HE-fed diabetics and asymptomatic rats, respectively. The last 2 values were significantly different (p less than 0.005). There were no other differences in plasma or tissue levels of vitamin E in these groups of animals. These findings suggest that high dietary vitamin E may decrease the incidence of diabetes in animals which are able to accumulate sufficient amounts of the vitamin in the thymus. Since the thymus plays a key role in the maturation of T cell populations, which appear to be altered in this disease, it seems possible that the protective effect may be exerted at this level.