A direct method has been described which makes possible a specific assay of progesterone in rat serum without extraction. Anti-progesterone serum was prepared in our laboratory by the immunization of three rabbits with 4-pregnen-3, 20-dione-3 CMO:BSA. This antiserum (Gunma OGP#1) displayed little or no cross reaction with 20 alpha-dihydroprogesterone (0.38%), pregnenolone (0.44%), 17 alpha hydroxypregnenolone (less than 0.1%), 20 beta hydroxyprogesterone (2.4%), 17 alpha hydroxyprogesterone (2.88%) or deoxycorticosterone (2.19%). The nonspecific inhibitory effect of serum was compensated for by adding progesterone-free serum to the standard curve tubes. The sensitivity of this assay was 1.1 pg/tube and serum progesterone could be measured by using as little as 1 microliter of serum. The working range of the standard curve was 1.25-2560 ng/ml. Under the conditions of this assay (1 microliter of serum per tube), interference from steroid binding proteins did not affect the sensitivity, precision or reliability of the assay. The intra-assay and inter-assay coefficients of variation were 5.5% and 8.7%, respectively, and the assay values correlated well with those obtained by the extraction method (R = 0.997, P less than 0.001). Analytical recovery indicates a close correlation between added and recovered progesterone concentrations (R = 0.992, P less than 0.001), and the recovery rate averaged 96%. Compared with the extraction method, the direct progesterone assay has the advantage of speed, precision and simplicity. The method described is particularly suitable for routine assays of progesterone in rat serum.