With the object of applying the van Slyke method to the study of the nutritive value of the protein in feeding-stuffs, we investigated first the extent of the error arising from the presence of interfering substances, and then searched for the simplest method to enable us to obviate, or at least to reduce, the chief source of error. Casein was hydrolyzed with or without the addition of the following substances: starch, sucrose, filter paper, olive oil, and agar-agar. The addition of fat influences the analytical results slightly, while the addition of carbohydrate interferes to a much greater extent with the determination of amino acids, by the method under consideration. These results coincide with those of YANAGAWA and NISHIMURA. Hence, before subjecting the sample to the van Slyke method, it is quite necessary to treat feeding-stuffs so that all the proteins separated from the interfering substances. By treating the sample in such a manner as to boil with water and precipitate with acetic acid, it was easily possible to free the proteins from most of non-protein nitrogenous substances and hotwater-soluble carbohydrates, of which the main parts of the carbohydrates in feeding-stuffs consist. Our experiment showed that the recovery of nitrogen in the form of coagulable nitrogen amounted in the cases of sardine muscle and soy-bean to 94-95% of the protein nitrogen after the Stutzer method, and to 91% of the protein nitrogen in the case of silk-worm pupa. Even following the above method, the fibre, which is contained only in the feeding-stuffs of plant origin, will remain in the coagulable protein part. Thereafter, we made further experiment, in which casein mixed with equal amount of fibre was analyzed on one hand by the original van Slyke method, and on the other hand by the Hamilton method, and by the another modified method in which analysis was made in the filtrate obtained after removal of the residue insoluble in 20% HCl. From the results obtained by these three methods, it was evident that there is no choice, on the whole, between them. Hence, we conclude that it is not necessary to treat the sample of feeding-stuffs with such a complex manner as HAMILTON recommends, if we analyze the isolated coagulable protein above-mentioned, which is free from most of the interfering substances.