1. All except two of 106 strains of hemolytic streptococci isolated from man, other animals, milk, and cheese have been classified into five groups, which bear a definite relationship to the sources of the cultures. These broad groups may be subdivided into specific types by methods discussed elsewhere. The specific group classification is made possible by employing two special reagents: (a) extracts prepared by treatment of the bacteria with hot hydrochloric acid, and (b) serum of animals immunized with formalinized cultures. This differentiation is not detected by the agglutination reaction. The grouping agrees with that described by other investigators on the basis of cultural and biochemical characteristics. 2. The group-specific substance present in strains of Group A has been identified chemically as carbohydrate in nature. The chemical composition of the specific substances upon which the specificity of the other groups depends has not been determined. It seems not unlikely, however, that all of them may belong in the general class of carbohydrates, each being chemically distinct and serologically specific in the individual groups.

The classification of group A streptococci into types depends upon specific proteins, the M antigens, characteristic for each type. These M antigens are associated with virulence, and streptococci which contain them give rise to antibodies which promote phagocytosis of streptococci of homologous type and lead to type-specific protection of infected animals. Anti-M antibodies are demonstrable by M-precipitin reactions, bactericidal tests, and active and passive protection tests in mice (1-8).In studying the M proteins, it was found that the protein antigen which served to identify strains designated by Griifith as type 28 differed from the type-specific antigens of all other types in that it was not inactivated by trypsin and several other proteolytic enzymes (9). Pepsin was the only enzyme found which destroyed this antigen. Further studies showed that strains containing this antigen did not represent a single type, and that the antibody which had been used as the basis for their classification into one type did not protect mice against infection with these strains. Since this trypsin-resistant protein, now designated as R antigen, lacked one of the most characteristic properties of type-specific M antigens, namely, the ability to stimulate protective antibodies, it was obviously not the type-specific substance. Additional investigation of these strains was therefore undertaken (10).Initial attempts to classify two of these R-containing strains yielded inconclusive results: mouse protection and precipitin tests for M antibody in antiserum prepared against each were not definitive but suggested that, although partial cross-protection occurred, one strain might belong to type 2 and the other to type 13.In order to clarify these relationships it was thought possible that bactericidal tests might reveal specific M antibodies more readily than the tests used previously. The bactericidal method depends upon the phagocytosis and destruction of streptococci by human leukocytes in the presence of M antibody. The results obtained have the same significance with regard to immunity as those 525 on

Rabbits immunized with whole streptococci of group B, type II, produce two immunologically distinct type-specific antibodies which are essentially equal by weight in protecting mice against infection with homologous type strains.The capsular antigen with which these antibodies react is a polysaccharide containing galactose, glucose, glucosamine, and a labile component which has not been chemically identified. Extraction of the bacteria with TCA yields this ‘complete’ antigen, whereas extraction with HCl yields a partial antigen without the labile component. This degraded antigen can also be derived from the TCA antigen by treating the latter with hot HCl, and is indistinguishable from that extracted directly from the bacteria with HCl.One of the antibodies produced, the TCA antibody, is directed against the labile component of the polysaccharide. The other, the HCl antibody, is directed against a β-d-galactoside determinant; and the precipitin reaction with this antibody is not masked in the ‘complete’ TCA antigen by the presence of the labile component.The group-specific polysaccharide, which is located in the cell-wall, is also extracted with either TCA or HCl but can be eliminated from the preparations by fractional precipitation with ethanol. Although it is known that the type-specific polysaccharide is located in the streptococcal capsule, it is not at present clear in what form this substance occurs in the living streptococcal cell. It may be present partially as the degraded HCl form, or possibly wholly as the intact TCA form. Further immunological and chemical studies of these type-specific polysaccharides are in progress, and will be presented in another communication.

Two methods of typing group A hemolytic streptococci have been described: slide agglutination and anti-M precipitin tests. The first, developed by Griffith (1), is more extensively used. I t has the distinct advantages of requiring small amounts of serum and culture, and of a quick reaction time when all conditions are favorable. It has certain disadvantages: The granular growth of some strains of streptococci is so marked that much subculturing is necessary before they are suitable for testing, and many days or weeks may elapse before the final result is available. Cross reactions are not uncommon, and make it impossible to differentiate certain types unless specially prepared sera are used which have been absorbed with regard to the antigenic relationships among types. This is due, in part, to the presence in group A streptococci of two antigens related to type specificity, so called M and T, which in a given microorganism do not always correspond (2, 3). Antibodies for one or both of these antigens have been used for the identification of types by the agglutination method. It is no~v evident that several types contain common or closely related T antigens, but that members of each type elaborate a distinct type-specific M substance (4). So far, we have only once encountered the antigenic combination consisting of two types with a common M substance and distinct T antigens. Probably a thorough understanding of the usual antigenic make-up of each type will eventually result in more satisfactory slide agglutination tests.Anti-M precipitin typing is more precise in that a single antigen-antibody system is involved, and there is relatively little crossing among types when

Under uniform diet conditions the normal bile fistula dog will eliminate pretty constant amounts of cholesterol—about 0.5 to 1.0 mg. cholesterol per kilo per 24 hours. Diets rich in cholesterol (egg yolk) will raise the cholesterol output in the bile but compared to the diet intake (1.5 gm. cholesterol) the output increase in the bile is trivial (5–15 mg.). Calves' brains in the diet are inert. Bile salt alone will raise the cholesterol output in the bile as much and often more than a cholesterol rich diet. Bile salt plus egg yolk plus whole bile give maximal output figures for bile cholesterol—60 mg. per 24 hours. Liver injury (chloroform) decreases both bile salt and cholesterol elimination in the bile. Blood destruction (hydrazine) fails to increase the bile cholesterol output and this eliminates the red cell stroma as an important contributing factor. Certain cholagogues (isatin and decholin) will increase the bile flow but cause no change in cholesterol elimination. The ratio of cholesterol to bile salt in the bile normally is about 1 to 100 but the bile salts are more labile in their fluctuations. The ratio is about reversed in the circulating blood plasma where the cholesterol is high (150–300 mg. per cent) and the bile salt concentration very low. Cholesterol runs so closely parallel to bile salt in the bile that one may feel confident of a physical relationship. In addition there is a suspicion that the bile cholesterol is in some obscure fashion linked with the physiological activity of hepatic epithelium.