288 PROTEIN POISONS 



that the poison does not come, solely at least, from the 

 protein used in the reinjection. Friedemann says that it is 

 generally held that the poison comes from the antigen, 

 but that this is pure hypothesis. He holds that the poison 

 may come from any one of the factors in the reaction 

 the anaphylactogen, the amboceptor, and the complement. 

 He holds that the minimum killing dose of the protein 

 on reinjection is so small that it cannot be supposed to 

 furnish a fatal quantity of the poison, and he thinks that the 

 ferment once set in action by the reinjection may go on and 

 digest the proteins of the animal body. Friedemann has 

 done most valuable work on metabolism in anaphylaxis, and 

 he holds that the increase in the nitrogen output is greater 

 than all of this element contained in the reinjection, and 

 therefore he thinks that the evidence that the whole of 

 the poison at least does not come from the foreign protein 

 is incontrovertible. He is undoubtedly right in his finding 

 that nitrogen metabolism in anaphylaxis is far beyond 

 that which can be accounted for by the nitrogen in the 

 foreign protein, and in this he has been confirmed by others. 

 Our own work 1 proves the same thing, but in our opinion 

 this does not show that the poison itself has any other 

 source than the protein of the reinjection. In the first 

 place, as we have seen, the minimum of the protein neces- 

 sary to produce anaphylactic shock is much greater than 

 that necessary to sensitize. Rosenau and Anderson sensi- 

 tized one guinea-pig with 0.000001 c.c. of horse serum, 

 and Besredka found sensitizing doses under 0.001 c.c. 

 uncertain, and he found the smallest killing dose to be 

 ^5- c.c. even when given intravenously. It will be seen 

 from these figures that there is a big difference between 

 the sensitizing and the killing dose. One-fortieth of a cubic 

 centimeter of horse serum contains about 2 mg. of protein. 

 We found in our work that serum albumin yields about 

 one-third its weight of poison, then 2 mg. would yield 

 0.66 mg., and the protein poison obtained by us, in a crude 



1 Jour. Amer. Med. Assoc., 1909. 



