IMMUNITY AND ANAPHYLAXIS 201 



the guinea-pig in 4 to 5 days. Piactically this was not 

 found to be so, but that a considerable excess of toxin 

 units is required to kill the guinea-pig, (stated by Ehrlich 

 to be 100 toxin units, and by others as intermediate 

 amounts). On this basis Ehrlich has built a whole structure 

 of epitoxoids, toxoids, prototoxoids, syntoxoids, and toxons, 

 designed to explain on the laws of chemical equivalence 

 and differences in chemical affinity, the apparent contra- 

 diction of these results. It is right to point out, however, 

 that the basis is not a chemical one. The death of a 

 guinea-pig in 4 to 5 days cannot be compared, though it 

 follows on an injection of toxin, to the appearance, more 

 or less immediately, of a precipitate in a liquid to which 

 some other chemical body has been added. Also, the 

 relationship between the mixture of toxin and antitoxin 

 which just causes no symptoms, and that mixture which 

 kills a guinea-pig in 4 to 5 days, is an arbitrary one, and 

 any apparent numerical relationship should be regarded as 

 fortuitous until other evidence shows that it is more than 

 that. Again, the admitted instability of the toxins, and 

 probably of antitoxins, even under every precaution, 

 renders the results equivocal. In brief, Ehrlich's theory 

 is that antitoxin has a valency of 200 for toxin, and that 

 some of the bonds of antitoxin can be satisfied by degenera- 

 tion products of toxin, prototoxoids, deuterotoxoids, 

 tritotoxins, of alpha and beta varieties, and by toxons, 

 which are not derived from toxin but are present in 

 the toxin fluid at first. [" Valency = 200 " cannot be 

 accepted in the chemical sense. It only means that, 

 starting with an amount of antitoxin which neutralized 

 100 toxin units elsewise defined, it was found that when 

 the conditions were changed, the antitoxin, in some way 

 or other, neutralized 200 of these same toxin units, or in 

 the experiment appeared to do so. If the amount of 

 antitoxin used at first had been the amount required to 

 neutralize 1 toxin unit (and von Behring advised the 

 larger quantity only for safety), the amount found in the 

 second experiment would have been 2 by inference. The 

 mixture of the toxin and antitoxin before injection may be 

 a factor of moment in regard to this change of valency.] 

 Ehrlich calls the quantity of toxin which just neutralizes 



