BACTERIOLYSIS AND ALLIED PHENOMENA 157 



components, and which acts hgemolytically on all red corpuscles. 

 It also acts much more quickly than ordinary solutions of venom 

 even when activated with lecithin, and it is very thermostable, 

 resisting boiling for six hours. It is evidently a definite chemical 

 substance and of great theoretical interest, as being apparently 

 an amboceptor-complement compound preparable in a pure state. 

 The analogy between the complements and the exotoxins leads 

 to the inquiry whether the production of an anticomplement occurs 

 when an active serum is injected into an animal of another species. 

 Ehrlich and Bordet both proved this to be the case. He injected 

 fresh horse serum into goats, and obtained a serum which would 

 prevent the normal complementing action of horse serum. 1 He 

 easily showed that this was not due to any action which it exerted 

 on the red corpuscles (rabbit's) themselves or on the amboceptor. 



FIG. 38. SHOWING THE EFFECT OF ADDING TO SENSITIZED RABBIT'S 

 CORPUSCLES (a, b) A MIXTURE OF FRESH HORSE SERUM (CONTAINING 

 COMPLEMENT, r), AND SERUM FROM A GOAT WHICH HAD BEEN INJECTED 

 WITH HORSE SERUM AND CONTAINED ANTICOMPLEMENT, d. 



The complement and anticomplement combine, and the corpuscles are 

 unaffected ; this is shown by the fact that they will dissolve on the 

 addition of fresh complement. 



It was clear, therefore, that it acted as an anticomplement. The 

 question now arose, Was this action exerted on the haptophore or 

 the zymophore group ? The following experiment shows that it 

 acts on the former : Sensitized red corpuscles were treated with a 

 mixture of normal serum and of serum containing anticomplement : 

 no solution took place. The corpuscles were then centrifugalized 

 off and fresh serum added : haemolysis occurred. Evidently the 

 anticomplement had not acted on the zymophoric group, for if it 

 had the haptophore group would have combined with the ambo- 

 ceptors of the sensitized corpuscles, and would have shielded 

 them from further action when fresh serum was added. What 



1 There is a possible fallacy here, to be mentioned under our description of 

 Bordet and Gengou's phenomenon. It may be simply a binding of the 

 complement in the precipitate. This is discussed later, and at present 

 Ehrlich's views will be set out as if they were definitely proved. 



