Some years ago it was shown by Avery (1) that the protective substances of antipneumococcus horse serum are intimately associated with the globulins. By means of fractionation with ammonium sulfate it was demonstrated that the antibodies are distributed among the eu-and pseudoglobulin constituents. The immune bodies were likewise found to be partially precipffated when antiserum is saturated with sodium chloride, or when the electrolytes, normally present in the serum, are removed by dialysis.Later methods for the isolation of pneumococcus antibodies were directed toward the dissociations of the immune precipitate obtained by the absorption of antiserum with pneumococci of the homologous type, or with aqueous extracts of the bacterial cells. The first of these studies was made by Gay and Chickering (2) who found that aqueous extracts of pneumococci, when added to homologous antiserum, yielded a precipitate which could be partially dissociated by means of dilute sodium carbonate. Such solutions contained not only type-specific agglutinins and precipitins, but protective antibodies as well. Later work by Huntoon (3) and his associates led to the belief that the immune substances were probably non-protein in nature.Further advances in the identification of the chemical nature of antibodies have been made as a result of the extensive investigations of Felton (4). It has been shown by him that when pneumococcus antiserum is diluted with distilled water, the major portion of the antibodies separates as an insoluble precipitate. Not only can the immune bodies be precipitated by this means, but they can be effectively separated from the serum by dilution with small quantities of ethyl alcohol (5). The conditions under which antibodies can best be separated from antiserum have been thoroughly investigated (6) and the product which has been isolated has been found to have properties which definitely characterize it as a globulin more basic than ordinary serum globulin (7).The further purification of the water-insoluble globulin bearing the immune bodies of pneumococcus antiserum has also been attempted by Felton (8). It has been shown that fractions of the protein, varying in their solubility in different 179 on