In the United States there are substantial geographic variations in selenium concentrations of foods, forage crops, and human blood. The human requirement for selenium Is not known, but selenium is known to be an essential trace element for the growth of quail,'chickens'9'and rats. 3 If dietary selenium or another trace element has an effect on public health, areas adequate or high in selenium or other trace elements may show a'human disease incidence or death rate different from those areas where selenium or another trace element is deficient or absent. Diets of torula yeast containing 1 ppm of sodium selenite significantly reduced4 the incidence of tumor in both the number of animals with 7, 12 dimethylbenz (7) anthracene-croton oil skin tumors (PUl. 05) and the number of animals with benzo (7) pyrene induced skin tumors (P< O. OOl). Clayton and Baumann'have observed a decrease in the number of diethylaminoazobenzene induced liver tumors in rats fed 5 ppm of sodium selenite. Topically administered sodium selenite4 has also dramatically reduced the number of skin tumors in animals (PG. 05) in five of the six tumor promotion experiments with DMBA-croton oil, DMBA-croton resin, or DMBA-phenol. The topical experiments were confirmed by Riley6 using DMBA and compound AI, the active principle of croton oil. Although the results could not be confirmed by the extensive testing by Harr et al.,'selenium at doses approaching chronic poisoning has been reported to produce some low grade tumors in rats." 9'

Because of the striking inhibitory effects of selenium in animals against a variety of carcinogens, and the possibility that selenium at higher levels might acfually be enhancing carcinogenesis, the object of this epidemiologic study is to extend these observations to determine whether or not