MECHANISM OF THE ACTION OF ANTIAMBOCEPTORS. 567 



ceptors also against the cytophile group. We have therefore ex- 

 amined another antiamboceptor serum, and compared its properties 

 with those of the antiserum made by injections of normal rabbit 

 serum. This serum, like the latter, was also obtained from a goat, 

 but instead of using normal rabbit serum for immunization the goat 

 had been treated with the serum of a rabbit previously immunized 

 with ox blood. Our experiments, however, did not permit of a 

 decision on this point. We are unable to say whether among the 

 antiamboceptors excited by the injections of the immune serum there 

 were any directed against the cytophile group. It is entirely con- 

 ceivable that, despite the presence of the cytophile group, these are 

 unable to exert any immunizing power, since the complementophile 

 groups invariably encounter the corresponding coufcter-group in the 

 organism and so are the only ones bound to the tissue receptors. In 

 that case previous to injection one would attempt to destroy the 

 complementophile group ( = cytophilic amboceptoids) or to neutralize 

 it by means of a suitable antibody. The decision of this question 

 must be left to further detailed investigations. 



In the course of our experiments we met with a very curious phe- 

 nomenon, one not only of some practical significance, but also of 

 considerable theoretical interest. Our experiment showed exactly 

 the opposite behavior which Bordet had found. That is to say, where 

 Eordet found that the antiserum acts as an antiamboceptor on the 

 amboceptor anchored to the cell, and that this action is overcome 

 by normal rabbit serum, one of our cases represents the reverse of 

 this. We see, therefore, that it can happen that the antiamboceptor 

 as such does not act, but requires the addition of normal rabbit serum 

 before exerting its action. We have constantly observed that in a 

 "curative" experiment, i.e., after a previous binding of amboceptor 

 and cell, large amounts of the antiserum produced by means of im- 

 mune serum were unable to prevent haemolysis. The following proto- 

 col may serve as an example: 



To each of a series of test-tubes, containing decreasing amounts 

 of the antiserum, 1 cc. of ox blood was added. This blood, after 

 having previously been sensitized with 0.003 cc. ( = 1J amboceptor 

 units) of an amboceptor obtained from a rabbit by immunization 

 with .ox blood, was freed from serum constituents by centrifuging 

 and then used in the test. After digesting the mixtures for half an 

 hour the blood-cells were centrifuged off and the sediments, to 

 which 0.1 cc. guinea-pig serum was added as complement, were 



