In a prospective study of about 22,000 men attending a screening centre, serum samples were collected and stored. The concentration of vitamin E (alpha-tocopherol) was measured in the stored serum samples from 271 men subsequently notified as having cancer and from 533 unaffected controls, matched for age, smoking history and duration of storage of the serum samples. The mean vitamin E level of the cancer subjects was not significantly different from that of their matched controls. The mean level in the cancer subjects who were diagnosed as having cancer before the elapse of one year from the date of blood collection was, however, significantly lower than the mean concentration of their matched controls (10.0 and 11.5 mgl-1 respectively, P = 0.003). For subjects whose cancers were diagnosed one or more years after blood collection the difference was not statistically significant either for all cancers or for cancers of six sites considered separately, viz. lung, colon and rectum, stomach, bladder, central nervous system and skin. The most likely explanation for these results is that the low vitamin E levels observed in these subjects were a metabolic consequence, rather than a precursor, of the cancer. This would explain, at least in part, the overall inverse association between serum vitamin E and risk of cancer observed in the published epidemiological studies on serum vitamin E and cancer.