Brain tumors from 64 patients were studied for the presence of hormone binding. Estradiol binding was detected in 34 patients: 13 of 21 meningiomas, 4 of 8 schwannomas, 1 oligodendroglioma, 1 of 5 benign gliomas, 4 of 5 malignant gliomas, 5 of 11 metastatic tumors, and 6 of 11 pediatric tumors: 1 medulloblastoma, 2 malignant ependymomas, 1 benign astrocytoma, 1 malignant sarcoma, and 1 malignant teratoma. Eleven patients were studied for progesterone binding, which was measurable in 7: 1 schwannoma, 3 meningiomas, 1 malignant sarcoma (pediatric group), 1 astrocytoma--gemistocytic (pediatric group), and 1 metastatic adenocarcinoma. There were 41 females and 23 males in the study. Fifteen females were premenopausal, 18 were postmenopausal, and 8 were in the pediatric group. Of the 34 tumors with measurable estradiol binding, 23 occurred in females. In the progesterone group, 4 of the 7 tumors with measurable binding activity were from female patients. In the pediatric group, estradiol binding was detected in 1 medulloblastoma, 2 malignant ependymomas, 1 malignant teratoma, 1 malignant sarcoma, and 1 astrocytoma. Five of the 6 pediatric tumors with estradiol binding were malignant, and both pediatric tumors with progesterone binding were also malignant. Of the 10 gliomas studied, 4 of the 5 malignant tumors had estradiol binding, whereas only 1 of the 5 benign tumors showed binding. Our studies with the pediatric tumors and the gliomas suggest that a relationship exists between the malignancy of the tumor and the presence of hormone binding. The ubiquitous nature of the presence of hormone binding is discussed, as is the possible correlation between age, sex, histological grade, and significance of hormone binding in brain tumors.