1. The absorption spectra of a number of proteins in the region 2500 to 3000 A. have been found to comprise from six to nine narrow bands. In consequence of variation in the relative intensity of these bands from protein to protein, the absorption curve has a characteristic configuration for each protein. 2. These bands correspond closely in position with the narrow bands which appear in the absorption spectra of tryptophan, tyrosin, and phenylalanine. Tryptophan and tyrosin each present three bands, phenylalanine shows nine. 3. The bands in the proteins are accordingly attributed to these amino acids. In the proteins the bands are displaced from the positions which they occupy in the uncombined amino acids, in most instances, by 10 to 35 A. toward longer wavelengths. 4. The absorption spectrum of Pneumococcus Type I antibody resembles that of normal pseudoglobulin but shows characteristic differences.

The reactions of hog, human, and horse blood group A substances with antibodies produced on injection of these substances into individuals of blood groups B and O have been studied by hemagglutination inhibition and by quantitative precipitin technics. Species differences in reactivity with antisera exist among these blood group A substances. Hog and human A substances are very closely related chemically as evidenced by their extensive cross-reaction while horse A substance crossreacts much less with antisera to both hog and human A substances. Analysis for hexosamine of specific precipitates of hog and human A substances and antibody to human A substance formed as a consequence of heterospecific pregnancy has shown that essentially all of the glucosamine of the blood group substance is precipitated.

1. Blood group substances have been prepared from human saliva, stomach, and amniotic fluid from individuals of blood group A(1) and A(2). Several of the saliva samples were obtained from individuals shown to be heterozygous, A(1)O. 2. The purified blood group A substances from human sources were similar in nitrogen, glucosamine, reducing sugar, and acetyl content. The A(1) and A(2) substances differed in optical rotation. All of the human A samples were levorotatory while those from hog stomach were dextrorotatory. 3. By two immunochemical criteria the various human preparations could be shown to fall into distinct groups, with respect to purity. The best products showed maximal activity and almost all of their glucosamine was specifically precipitable by anti-A. These samples of human A substance were only about one-half as effective in precipitating antibody to hog A substance formed in man as was homologous hog A substance although the same total amount of antibody was precipitable by excess of either antigen. 4. Human blood group A(1) substance was found to be antigenic in individuals of blood groups B and O but was not as good an antigen as hog A substance.

1. A picture resembling acute disseminated encephalomyelitis in the human being has been regularly and rapidly produced in rhesus monkeys by injection of emulsions of adult rabbit and monkey brain administered with adjuvants. 2. No lesions of the central nervous system resulted from injection of similar emulsions of fetal rabbit brain or adult rabbit lung. 3. A description of the gross and histological findings in the central nervous system is given and compared with features of human demyelinating disease. 4. The experimental findings are in accord with the hypothesis that antibody to the injected brain emulsion reacts with the tissues of the nervous system of the animal to produce the pathological changes.

1. Studies on a number of individual hog stomachs have shown that substances with blood group A activity cannot be obtained from all hogs. Of ten stomachs studied, only seven yielded products with blood group A activity. All ten purified preparations, however, showed identical properties with respect to nitrogen, reducing sugar, glucosamine, acetyl, and relative viscosity. Six of the seven active samples were of equal potency in precipitating anti-A; the seventh was slightly less active. 2. Preparations from random pools of hog stomachs, although possessing the same analytical properties, were of lower activity than those from individual active stomachs as determined by the microquantitative precipitin method. 3. An immunochemical method for estimating the absolute purity of the blood group A substance by determining the proportion of its glucosamine precipitated by excess anti-A was developed. Values of about 84 per cent for the purity of six of the seven purified active preparations from the individual hog stomachs were obtained. 4. The inactive products, unlike the active ones, did not stimulate the production of anti-A on injection into human beings and did not precipitate anti-A or inhibit hemagglutination of A erythrocytes by anti-A.

1. The microquantitative precipitin method can be used to compare the relative activity of different preparations of the blood group A substance from hog stomachs and to study the effect of chemical treatment upon its stability. 2. With samples of about 25 microg. antibody nitrogen, an error of +/-1.7 microg. antibody nitrogen will result in an error of +/-12 per cent in the estimation of the amount of A substance. 3. The blood group A substance showed no significant loss of activity at 37 degrees C. after 48 hours at pH 1.07 to 10.7 or after 2 hours at 100 degrees C. over a pH range from 2.97 to 7.58. Exposure at 100 degrees C. at pH 1.03 or at 9.03 or higher resulted in loss of activity. Parallel results were obtained by the hemagglutination inhibition and quantitative precipitin methods. 4. The solubility of specific precipitates of the blood group A substance from hog stomach and its homologous antibody formed in man was found to be about 1.6 microg. antibody N/ml. 5. A comparison is given of the chemical properties and activity of blood group A substances obtained by several procedures from pools of hog stomachs.

1. With purified blood group A and B specific substances from the gastric mucin of pigs precipitin reactions can be obtained in sera containing homologous isoagglutinins. By microquantitative precipitin technics the anti-A and anti-B content of normal and immune isoagglutinin-containing sera can be estimated on a weight basis. 2. There was a general correlation between antibody N precipitable from the sera of groups O and B by A substance and the agglutinin titer for A erythrocytes before and after immunization with blood group substances. 3. The reaction between A substance and its homologous isoagglutinin present in the sera of immunized individuals was of the same type as in other antigen-antibody systems.