446 H^EMOLYSINS AND ALLIED BODIES 



some normal serum, i.e. containing complement : no haemolysis 

 resulted. The sediment (erythrocytes) was suspended in isotonic 

 saline and some normal serum i.e. containing complement 

 added : complete haemolysis resulted in two hours. The erythro- 

 cytes had therefore united with the amboceptors. It was further 

 shown that this union was not merely a mechanical adherence ; 

 for, repeatedly washing the sediment with isotonic saline did not 

 influence the above result, the washed erythrocyte-amboceptor 

 compound haemolysing, when complement was added, as readily 

 as when it had not been washed. Moreover, the union was a 

 specific chemical one, for a repetition of the experiment with any 

 erythrocytes other than those of a sheep gave a sediment con- 

 taining no amboceptors (i.e. it did not haemolyse when comple- 

 ment was added to it). 



II. Does the complement also become attached to the erythrocytes, or 

 can it only become attached to erythrocytes already united with 

 amboceptors ? 



Some serum of a normal goat was mixed with sheep's erythro- 

 cytes, which had been washed free of adherent serum, and, after 

 some time, the mixture was centrifugalised. The corpuscular 

 sediment, after washing, was found to contain no complement, for 

 the addition to it of inactivated haemolytic serum (i.e. containing 

 only amboceptors) gave no haemolysis. The complement cannot 

 therefore unite directly with erythrocytes. 



III. How, then, does the complement act on the erythrocytes? 



We have seen that it cannot combine with erythrocytes 

 directly, whereas the amboceptors can. From which fact we 

 might at once assume that during haemolysis the complement 

 acts on the erythrocyte through the intermediation of the ambo- 

 ceptor. To ascertain if this be so, Ehrlich and Morgenroth 

 studied the behaviour of complement and amboceptor on erythro- 

 cytes when both were present in the mixture, (a) To prevent 

 the rapid haemolysis which would follow at ordinary tempera- 

 ture in this case, the mixture was cooled to 0-3 C., at which 

 temperature haemolysis is very slow. After leaving the mixture 

 for some time at this temperature, it was centrifugalised, and 

 sediment and supernatant fluid independently examined for the 



