The present paper summarizes the experience of the authors in the setting up of the radioimmunoassay (RIA) for human follicle simulating hormone (H-FSH), with the purified antigen for radioiodination, the F-FSH standard and the specific antibodies kindly donated by the National Pituitary Agency of the National Institutes for Arthritis, Metabolic and Digestive Diseases of the National Institutes of Health, Bethesda, MD. U.S.A. The conditions for the RIA have modified somewhat and simplified with respect to the suggested instructions accompanying the reagents. Thus, the amount of Chloramine T and the time of exposure of the labeled H-FSH (H-FSH) has been studied. It is always purified on Sephadex G-100 immediately before addition to the RIA and in this manner it may be used for up to 2 month after labeling when kept at --20 degrees C. Curves obtained at different dilutions of the H-FSH Standard, carried out with phosphosaline buffer, pH 7.4-7.8 (PBS) containing 1 % human serum albumin, or with horse plasma, of with PBS containing 0,25 % serum from non-immune rabbits (RIA Buffer) have been compared iwth those abtained by serial dilutions of sera from post-menopausal with these diluents. The most consistent results were obtained with the RIA buffer as diluent. The redisual error was smaller, and serial of dilution curves of the H-FSH standard were parallel to those of plasma and acetone precipitates of urine from post-menopausal women. Parallelism was not god using those serum. Results using PBS contianing human seum albumin were poor. PBS containing bovine serum albumin was avoided as some batches were found to interfere with the binding of the F-FSH to the antibody. The stability of the different dilutions of the H-FSH standard prepared in RIA buffer was tested. It was found that the standard curves could be prepared, pipetted into the RIA tubes and kept ready, frozen at --20 degrees C for one to two months. This shortens the actual setting up of a given RIA and decreases interassay variation of results. Parallelism of the H-FSH standard curve with serial dilutions (in RIA-buffer) of sera from women on the day of the preovulatory was confirmed. The data obtained in men and women, during stimulation with LH-RH are also given. No cross reactivity was found the HCG or sera from women, in agreement with the fact that the antiserum had been absorbed with HCG. There is, however, a considerable degree of cross reactivity with H-TSH; Thus, sera containing 15 muU/ml H-TSH or more, would give false H-FSH results. Such H-TSH values are not only found in hypothyroid patients, but might be reached during TRH responses to TRH.