Mounting laboratory and epidemiologic evidence suggests that selenium may be important in the etiology of both cancer and heart disease. We explored the use of hair and nails as indicators of selenium intake by measuring their selenium levels using neutron activation analysis, a highly sensitive and precise nondestructive technique. Levels in duplicate samples of nails, hair, and blood were all reasonably reproducible. However, selenium-containing shampoos severely contaminated some of the hair specimens, suggesting that use of hair in epidemiologic studies could be misleading. The mean selenium level in toenails from South Dakota (a known high selenium area) was 1.17 ppm (1SE = 0.09). This was significantly higher than mean levels from Boston and Georgia (medium selenium intake area) of 0.74 ppm (0.04) and 0.81 (0.03), respectively. The mean selenium level in toenails from New Zealand (low selenium area) was 0.26 (0.02) and these levels did not overlap those of other areas. The South Dakota specimens showed marked familial aggregation, probably reflecting dietary differences. Since toenails vary in length, clippings from different toes represent different time periods of formation; clippings from all ten toes reflect selenium levels integrated over an extended period. As toenails are easily collected, transported, and stored, and reflect longterm intake, they can be useful in epidemiologic studies of selenium and chronic disease.