Abstract Data on 451 trials of endocrine therapy in 366 patients from 8 cities are reported. The relative odds relating the results of the ER assay to the probability of tumor regression following endocrine therapy, when adjusted for hospital, was estimated at 15.47, a highly significant result. For these data the dominant site of metastasis was not important in predicting response to endocrine therapy, but hospital, disease free interval, and the site of visceral metastasis were found to increase our ability to predict even when the results of the ER assays were known. The clinical variables analyzed in the report do not allow us to explain entirely why some 41% of patients with ER+ tumors failed to regress following endocrine therapy, but in one subgroup of 71 patients the regression rate for ER+ tumors was 78.9%. The response to a first course of endocrine therapy improved predictions based on the ER assay alone concerning response to a second course of endocrine therapy.