Carcinomas of the breast from 352 women were assayed for binding of tritiated estradiol by tumor cytosol with dextran--charcoal adsorption, saturation analysis, and two-point Scatchard plots; the level of saturable binding defined a cytosol as positive or negative for estrogen receptor. Valid assays were obtained on specimens as small as 120 mg. Assays of replicate samples of a cytosol were more reproducible than assays of replicate samples of the tumor itself. Occasional disparity of results between a primary mammary carcinoma and its axillary metastases could be related to differences in tumor cellularity. Saturable binding consistent with the presence of estrogen receptor was found in 59% of 305 primary carcinomas and in 57% of 47 metastatic or recurrent carcinomas. There was a significant negative correlation between the patient's age and saturable estrogen binding in the tumor. Serum estradiol levels of less than 250 pg/ml appeared to have a negligible effect on estrogen receptor content. A small subgroup of high-binding carcinomas had high dissociation constants, but the significance of this observation is not clear.