We have studied 29 patients from whom two or more spatially and temporally separate samples of breast carcinoma were obtained for ERP analysis. Differences in ERP were obtained in 24% of all cases studied. The greatest degree of variation was found when comparing a primary tumor with a subsequent metastasis (38%). Among eight patients with ERP negative (-) primary tumors, six had ERP (-) and two ERP positive (+) metastases. Among 11 patients with ERP (+) primaries eight had ERP (+) and three ERP (-) metastases. Variations were found in 20% of multiple nonsimultaneous metastases from individual patients. The difference in average interval between cases that did not show variation (9.6 months) and those that did vary (19.3 months) suggests that variation may be more likely to occur with later metastases. However, there was considerable overlap in the range of intervals. These findings underscore the need to biopsy readily accessible recurrences rather than to rely on the results of prior specimens. There was no apparent consistent relationship of these variations to age, site of specimen, interval between specimens or the histologic features of the tissues submitted for analysis. Two patients with an ERP (+) primary tumor had ERP (-) metastases after interval therapy. Similar variations were observed in patients who did not have interval therapy. The data do not permit any definite conclusions as to the effect of chemotherapy or radiation on the ERP properties of mammary carcinoma. Cancer 39 : 2 194-2200, 1977. STROGEN RECEPTOR PROTEIN ASSAY OF E breast carcinoma provides a helpful guide for predicting response to hormonal therapy in patients with metastatic breast carcinoma. l7 Nearly two-thirds of patients with an ERP positive tumor specimen have been reported to respond favorably while less than 10% of patients with ERP negative tumors experienced similar improvement. la Failure to obtain higher response rates in favorable cases could reflect differences in the ERP properties among tumor cells, 24332 the importance of other hormonal receptors , 13915 or other variations in metabolic