506 STl I'll s IN IMMUNT] 5 . 



which had been affected by the agglutinin, act subsequently pre- 

 cisely like inert chemical particles. In other words, bacteria, on 

 uniting with the agglutinin, give a coagulum which is flocculable by 

 electrolytes. But could we not go further and say that any easily 

 COagulable substance will give the same results? Should we not 

 obtain an agglutinating serum following the injection, not of formed 

 elements like bacteria, but of particles of amorphous organic sub- 

 stances? Milk casein, which occurs in a colloidal condition in the 

 form of extremely minute particles, is suitable to prove this point. 

 Animals that were immunized against milk gave a serum which 

 precipitated casein and consequently agreed with the observations 

 which Tchistowitch had just made on the formation of a precipitate 

 on adding eel serum to its antitoxin. 



The flocculation of milk casein by the corresponding immune 

 serum established a new connection between the phenomena of 

 agglutination and coagulation which Duclaux had previously 

 asserted to be closely related. Adsorption of the agglutinin modi- 

 fies bacteria in respect to their properties of adhesion with the 

 surrounding fluid, and this modification is evident by their sus- 

 ceptibility to electrolytes. But how does the agglutinin unite? We 

 might conceive of a single molecular contact, of an adsorption of the 

 antibody by the antigen similar to the adsorption of anilin dye by 

 filter paper. I tested antibodies from this standpoint, and in par- 

 ticular did the experiment which demonstrated that a given dose of 

 hemolytic serum destroys variable amounts of corpuscles in accord- 

 ance with their addition in a single or in several doses. It would 

 take me too far afield to discuss this experiment in greater detail, 

 an experiment which, as we know, has been employed by various 

 investigators with similar results, particularly in dealing with 

 antitoxins. Such are the purely experimental results which first 

 suggested to me the ideas that I have since defended concerning 

 the mode of action of antibodies, without being guided by any 

 theoretical leaning in one direction more than another. 



When we have hemolytic sera which are evidently toxic, it is 

 quite natural to attempt to obtain antitoxins to them; the effect of 

 such an antiserum on the hemolytic serum can then be split up into 

 several factors and it is found to contain both an antisensitizer and 

 an anti-alexin. From another standpoint, inasmuch as the alexin 



