STUDIES ON HJEMOLYSINS. 105 



This can only be explained by assuming that only part of the total 

 number of immune bodies find fitting complements hi goat serum, 

 and that this partial number varies, but is always less than the number 

 of immune bodies activated by guinea-pig serum. The diagram 

 presented below will best make this relation clear. 



We have made a further series of experiments in order to com- 

 plete these studies, and have discovered that our anti-immune body 

 also protected goat blood-cells against the immune body derived 

 from a rabbit treated with ox blood. This of course is quite natural, 

 for we have already shown that this action on a strange species of 

 blood depends on a concordance of certain haptophore groups. 

 Similarly, the anti-immune body protects ox blood-cells against the 

 immune body of a rabbit immunized with goat blood. 



These experiments lead us to the following conclusions: The 

 anti-immune body derived by injecting goats with immune bodies of 

 rabbits is not a simple uniform [einheitlich] substance, but is made up 

 of a whole series of partial immune bodies. In the ox blood used to 

 immunize the rabbits we have already distinguished two main portions 

 of receptors to which again two main portions of the resulting immune 

 body correspond. Each of the latter portions in all probability con- 

 tains a host of partial immune bodies, and we must assume that f 

 corresponding to this, the anti-immune bodies also possess a complex 

 constitution. 



In the following diagram it is not sought to express that the 

 immune bodies which can be activated by guinea-pig serum are all 

 identical. On the contrary each group may represent a different 

 kind of immune body. 



We have seen that there is a great difference between the dose 

 of immune body which is completely neutralized by a certain amount 

 of anti-immune body, and that which in the presence of anti-immune 

 body causes complete solution. This can be understood when we 

 recall the above-mentioned distribution of partial immune bodies, 

 and examine the diagram, Fig. 2. 



In order to choose a simple illustration let us assume that, corre- 

 sponding to the diagram, the immune serum of the rabbit immu- 

 nized with ox blood contains only two different types of immune 

 bodies and these, furthermore, in unequal amounts. Let the main 

 portion be represented by immune body type a, which is activated 

 by a particular complement present in the animaPs own (rabbit's) 

 serum. Further, let the second portion, present in much less amount, 



