Since the introduction of gas-liquid chromatography by Janes and Martin for the qualitaive assays of various chemical substances, this method proved to be an excellent one for isolation and identification of a variety of oily liquids as well as many other substances. The gas-liquid chromatography has been used also in the analysis of steroids affording many satisfactory results. However, this method has one great disadvantage in the when a minute quantity of urinary 17-KS (ketosteroids) is placed in liquid form it is not possible to avoid the loss of test material other than in the first trial, and also the peak of the solvent often interferes with the peak of the steroid making it difficult to make an accurate chart of the mixture of steroids of varying concentrations. Therefore, the authors devised a new gas-liquid chromatography in which the steroids are first adsorbed into a stainless steel gauze and then inserted into the specimen chamber, and we find that this eliminates the appearance of solvent peaks, as observable in the routine gas-liquid chromatography. By this method we separated and analyzed seven fractions of urinary 17-KS; namely, androsterone, etiocholanolone, DHA, 11-OH-adrosterone, 11-OH-etiocholanolone, 11-oxo-androsterone and 11-oxo-etiocholanolone. The apparatus used was the super-sensitive gas chromatograph with GC-1B solid insertion attachment and hydrogen flame ionization detector of Shimazu Mfg. Co. For the column we used 2% NGS (Shimazu product). The test material was a part of the 24-hour pooled urines. The urine was hydrolyzed by solvolysis and β-glucuronidase (the twostep hydrolysis), and the residue (neutral steroids) was sebarated into 17-KS and corticosteroids by Florisil column chromatography. Further, the 17-Ketosteroid fraction was separated into 11-deoxy-17-KS, 11-oxy-17-KS and Pd fraction by the alumina column chromatography. This chromatography serves as the prechromatography of the 2% NGS' gas-lipuid chromatography by which we can isolate DHA from 11-OH-androsterone as well as androsterone from Pd. With the fractions isolated in this manner we esperified them in the presence of trimethylsilyl ester, and we studied the ratios of anderosterone, etiocholanolone, DHA, as well as the ratio of 11-deoxy-17-KS to 11-oxy-17-KS in the 7 fractions isolated from the urines of female with normal menstrual cycle, normal adult males, and females with endocrine disorders. The results of our study may briefly be summarized as follows. In the urine of females of normal menstrual cycle adrosterone, etiochlanolone and DHA of 17-Ketosteroids tend to be higher before the ovulation and during the corpus luteum period. In the urine of normal adult male the pattern of 17-KS tended to show a high androsterone level and high androsterone ratio. In the case of Hirstism a marked increase of 11-deoxy-17-KS was observed, and the aderosterone ratio tended to show a similar value as in the male. In the cases of testicular feminization syndrome the characteristic feature was the fall in the aderosterone ratio. In the ACTH loading test. 11-deoxy-17-KS responded in the order of DHA>etiocholanolone>androsterone. Also 11-oxy-17-KS responded well. In the dexamethasone inhibition test of 11-deoxy-17-KS; DHA was inhibited most markedly, while 11-deoxy-17-KS were all inhibited uniformly.