An inverse relationship exists between dietary selenium (Se) concentrations and the incidence of human breast cancer. The addition of Se to the diet has been shown to decrease the incidence of spontaneous murine mammary tumors. We compared the serum Se concentrations in breast cancer patients with those of women without breast cancer. Serum was collected from 35 women with breast cancer. Nineteen of these women had infiltrating ductal carcinoma and two had Paget disease of the nipple. Nine women had lymph nodal metastases at the time of mastectomy, four had definite evidence of metastatic disease when the blood samples were drawn, and the disease process of one patient was unclassified. Samples from 27 women known to be free of breast cancer were used as controls. The difference noted between the mean serum Se concentrations of breast cancer patients and controls were found to be significant.