1 6 Society for Experimental Biology and Medicine. 



tigation indicated the unreliability of Pick's differentiation, and 

 attention was accordingly directed to the actual possibility and 

 practicability of distinguishing between anti-bodies by fractionation 

 of the globulin. The availability of poly-agglutinative sera for 

 the work gave a chance for making numerous and extended obser- 

 vations of the distribution of these anti-bodies in the fractions. 



It was found repeatedly in experiments with rabbit and goat sera 

 that the agglutinins for the dysentery group of organisms (Flexner 

 Manila and Shiga), typhoid, coli and cholera, were not confined 

 to either the pseudoglobulin or the washed [with 3.4 sat. (NH,) 2 S0 4 

 solution] euglobulin fractions ; they were either split by the 

 fractioning, the major portion occurring in the pseudoglobulin, or 

 almost the entire amount of the agglutinating substances recovered 

 were in this higher fraction in the original quantitative proportion 

 to one another. With anti-dysentery horse serum, the dysentery 

 (Shiga and Flexner) and coli agglutinins were fairly quantitatively 

 though not qualitatively split between the pseudo- and euglobulin 

 fractions, the latter containing the lesser amount. With an anti- 

 cholera and anti-typhoid horse serum, the pseudoglobulin (two 

 experiments) and also the filtrates from two additional 3.6 and 3.8 

 saturation precipitations contained the bulk of the agglutinins. 

 In subsequent experiments with sera from other bleedings as well 

 as with the sera used above, the typhoid agglutinin was divided 

 between the two fractions with a somewhat larger proportion occur- 

 ring in the pseudoglobulin. 



It is apparently difficult to control all the conditions under 

 which experiments of this type are made ; absolutely constant re- 

 sults at times cannot be obtained on successive repetition of the work. 



The results of exhaustion experiments on the two globulin frac- 

 tions were the same as those that would be obtained in the use of 

 the native serum, and failed to give any reason for believing that 

 we were dealing with a separation of group and specific agglu- 

 tinins through fractioning. 



Precipitation of anti-diphtheria goat serum showed that about 

 half the antitoxin remained in the pseudoglobulin ; practically none 

 was found in the euglobulin while the 3.4 saturated (NH 4 ) 2 S0 4 

 solution washings contained the balance. 



The results of the work thus far accomplished have demon- 



