Abstract. In prostatic cytosol DHT1 is metabolized to 5α-androstane-3α (or β), 17α-diols with a half life of 2 h even at 4°C. Thus, [3H]DHT appears to be a poor marker for a quantitative assessment of androgen receptors (AR). Methyltrienolone (R1881) seems to be advantageous as it is not metabolized. However, because of considerable binding to progestin receptors, assays using [3H]R1881 are not specific for AR in tissues containing progestin receptors. We, therefore, developed a specific assay for AR using [3H]DHT (14 nm) as marker, where metabolism of DHT is prevented by pre-incubation with NAD+-nucleosidase. The [3H]DHT-receptor complex is separated from free, SHBG-bound and unspecifically bound [3H]DHT by agar gel electrophoresis. The binding sites of high affinity and low capacity are characterized by suppression with unlabelled R1881 (2 μm) in a parallel assay. Under these conditions DHT and R1881 appear to have the same kinetics of association and dissociation. Weighted non-linear regression analysis of specific binding capacity at various ligand concentrations reveals that in rat prostatic cytosol the affinity of DHT (Kd = 0.405 ± 0.0839 nm) is significantly higher (P < 0.01) than that of R1881 (Kd = 1.25 ± 0.271 nm).