Evidence is presented for a unified model for the reaction of antiestrogens with estrogen receptors that explains much of the unusual pharmacology of these clinically important agents. Agonist activity results from occupancy of the estradiol-binding (primary) site in the receptor and antagonism from the additional interaction with a secondary locus not recognized by hormone. In the case of type I antiestrogens, such as tamoxifen, this is weaker than primary site binding, so these substances are agonists at low concentrations and antagonists at higher levels. With type II antiestrogens, affinities for both sites are comparable, so one never has the agonist situation (only primary site occupied), and these agents are pure antagonists. Reproductive tissues of the mouse and guinea pig contain a small macromolecule that binds hydroxytamoxifen, but not estradiol, keeping the free antiestrogen concentration below that required for secondary binding. Thus, in these species, tamoxifen is a pure agonist.