c. The ninhydrin test. To ten cubic centimeters of 

 antigen was added 0.2 cc. of a one per cent aqueous 



solution of ninhydrin (triketophydridene hydrate). The 

 mixture was boiled for exactly one minute after the 

 appearance of the first bubbles on the side of the tube. 

 No color change could be detected, even after cooling 

 and standing for four hours. This indicates the absence 

 of alplia amino acids. 



d. The cyanide test. About three cubic centimeters 

 of the antigen was evaporated, and the residue fused with 

 metallic sodium. The fused mass was placed in a small 

 amount of distilled water, boiled and filtered. To the 

 filtrate was added a few drops of fenous sulphate, fenie 

 chloride and hydrochloric acid. Absence of a blue colora 

 tion indicated an absence of nitrogen. 



It is evident that we are dealing with a nitrogen- 

 free substance. 



The original bacterial cells, after being doubly ex- 

 tracted with ether, were then suspended in 500 cc. of 

 sterile physiological sodium chloride solution. For 

 preservation, sufficient carbolic acid was added to make 

 a O.5$ solution. The greater part of the solids in this 

 mass are supposedly bacterial proteins and for experi- 

 mental purposes the mass is called the protein antigen. 

 Obviously it is a bacterial antigen minus the fats. 



For use as a control there was next prepared a 

 suspension in physiological saline, of a twenty-four hour 

 agar culture of the same strain of B. Coli, This was 

 made up with 0.5$ phenol and it constitutes the B. coli 

 of antigen. 



We now have three distinct antigens: 



1. Sodium salts of the fatty acids of B. coli. 



2. B. coli after being extracted with ether. 



3. A simple suspension of B. coli. 



It is the purpose now to determine whether or not 

 these antigens are true antigens; i.e., whether or not 

 they give proper antigenic reactions "in vitro", "in 

 vivo", or in both. To this end the antigens were inject- 

 ed into suitable animals, at certain intervals, in order 

 to obtain antibodies. 



The fatty antigen was diluted with physiological 

 salt solution, approximately one part of the alcoholic 

 solution to five parts of the saline. This gives a very 



