98 INFECTION AND RESISTANCE 



poids are capable of binding antibodies in miro, and that, in nervous 

 tissues, toxin fixation was in some way related to the richness in 

 lipoids of these structures. Bang and Forsmann 71 accordingly 

 treated animals with ether extracts of red blood cells claiming that 

 this resulted in the production of hemolysins. And these results 

 have been confirmed by Landsteiner and Dautwitz. 72 The latter, 

 however, suggest that the hemolysin production may have been in- 

 duced, not by the lipoidal substances in solution, but by other anti- 

 genie substances which had gone into colloidal suspension in the 

 ether extracts. Much similar research on the antigenic nature of 

 lipoids has been done, but, after reviewing this very thoroughly, 

 Landsteiner comes to the conclusion that no definite proof of the anti- 

 genic nature of any pure lipoid has so far been presented. The prob- 

 lem is experimentally complicated by the fact that, as Landsteiner 7a 

 suggests, the antigen may often be present as a lipoid-protein com- 

 bination, and as such go into solution or fine emulsion in the organic 

 solvents ; also the lipoids possess the curious property of altering the 

 solubilities of proteins and other substances by their presence. 



Summarizing our present knowledge of the chemical nature of 

 antigens, then, we must conclude that, with the exception of Ford's 

 glucosids, no protein-free antigens have been thus far demonstrated. 



In the light of this fact it is all the more remarkable that antigen- 

 antibody reactions are specific. For we possess no chemical methods 

 by which one variety of protein can be distinguished from another. 

 And yet the serum antibodies produced with each species of bacteria 

 react with this species only and the hemolysins, agglutinins, or 

 precipitins produced by the injection of bacterial, cellular, or serum 

 proteins react respectively only with the particular variety employed 

 in their production. This indicates that each of these antigens of 

 almost unlimited number must possess a chemical structure indi- 

 vidually characteristic and different from all the others. It is by 

 means of the biological reactions, indeed, that we can detect protein 

 in dilutions far beyond the reaction-sensitiveness of chemical tests 

 and can distinguish between varieties of protein when the chemical 

 methods will indicate only protein in general. Our knowledge of 

 the chemical constitution of protein has not yet advanced to a point 

 at which specificity can be based upon definite variations of chemical 

 structure, and the complexity of the problem is such that it does not 

 seem likely that we can hope in the near future to attain such knowl- 

 edge. We can merely accept it as a fact that the antibody produced 

 with one protein differs materially from that produced with another, 



71 Bang and Forsmann. Hofm. Beitr., 1906; Centralbl f. Bakt., 40, 1906. 



72 Landsteiner and Dautwitz. Hofm. Beitr., 9, 1907. 



73 Landsteiner. "Wirken Lipoide als Antigene?" Weichardt's Jahresbe- 

 richt, Vol. 6, 1910. 



