Data derived from a correlated morphological and biochemical study suggest the following: (a) estradiol-17~, diethylstilbestrol, the estrogen antagonists nafoxidine (Upjohn 11,100), and Parke Davis CI628 induce synthesis of an endogenous peroxidase in the epithelium of target tissues like the vagina, the cervix, the uterus, and in the acinar cells of the estrogen-dependent rat mammary tumor; (b) peroxidase is a "specific" secretory protein of the estrogen-sensitized uterine endometrium; (c) peroxidase synthesis is not a nonspecific response to steroid hormone action, since progesterone and testosterone do not induce its synthesis; (d) endogenous peroxidase is a possible diagnostic protein for the detection of estrogen-dependent growing tissues, including breast cancer; (e) movement of exogenous horseradish peroxidase from the interstitium to the uterine lumina is restricted by tight junctions located at the apices of epithelial cells. Estrogen and antagonists do not appear to influence the transepithelial movement of exogenous peroxidase into the lumen.The uterus, cervix, vagina, and mammary gland of mammals require estrogen for their postpubertal trophic development and for the maintenance of their differentiated state. Indeed, the presence of estrogen-binding proteins (estrogen receptors or estrophiles, Jensen et al., 1966 a) in these tissues and of estrogen dependency of rat and human mammary carcinoma are well documented (Huggins et al., 1959; Jensen et al., 1966 b;Heuson et al., 1972).In the course of our investigation it was discovered that tissues displaying growth dependency on estrogen synthesize and secrete "specific marker proteins" into their lumina. These results show (a) that synthesis of peroxidase ~ (or a hemoprotein with peroxidatic activity) is a consistent marker for estrogen-dependent tissues (carcinogen-induced rat mammary carcinoma, cervix, vagina, and uterus), (b) estrogen antagonists (Upjohn 11,100; ~Throughout this paper, the DAB-positive reaction product will be referred to as peroxidase, since uterine peroxidase is a well-documented enzyme (see text) and since the histochemical medium employed is designed to demonstrate peroxidase activity. It is of course possible that what is being visualized in either a hemeprotein, which may possess peroxidase activity, or several substances with peroxidasic activity. The biochemical characterization of this enzyme as a peroxidase is still very limited.