Androgen has long been known to act on the brain to modify behavior and other brain functions. In the past, two methods have been used to characterize the putative receptors which mediate these actions. Autoradiography has been used to map and identify androgen binding neurons. Binding studies have been conducted to quantify and characterize the system(s). The resultant data are discordant and a new model is proposed to resolve the apparent differences. It is proposed that there are three categories of receptors for androgen in the brain. One receptor preferentially binds testosterone and a second one preferentially binds DHT. Both of these receptors are in equilibrium between nucleus and cytoplasm according to the free water content of the compartments. Both of these receptors can be activated and transformed by steroid and thus concentrate in the nucleus. It is proposed that a third receptor binds both steroids with the same relative affinity. However, this third receptor can only be activated but not transformed (i.e. it does not concentrate in the nucleus). The proposed system implies that testosterone acts on a few discrete populations of neurons in the brain while DHT has a very diffuse action on the central nervous system.