8 Prof. B. Muir and Mr. C. H. Browning. [May 17, 



studied is always several times that of guinea-pig's complement. This 

 may be because there are fewer complement molecules in a given 

 amount of rabbit's serum or because the xymotoxic group of the 

 rabbit's complement is less active than that of the guinea-pig's 

 complement. In the former case a hsemolytic dose of rabbit's comple- 

 ment will prevent the combination of a haemolytic dose of guinea-pig's 

 complement; in the latter it will prevent the combination of more. 

 Our experiments are not sufficiently extensive to give a definite state- 

 ment on this point, especially in view of the fact that during the pro- 

 gress of an experiment the hsemolytic action of complement may 

 diminish, and this change in the value does not always occur in 

 the two complements in the same proportion. We may say, however, 

 that we have obtained in several instances a correspondence between 

 the combining and the haemolytic ratio, i.e., a haemolytic dose of guinea- 

 pig's complement keeps out a dose of rabbit's, and vice versA. We 

 are inclined to think that this may be the rule in the case before us, 

 and that probably the divergences in these ratios which we have also 

 met with are the result of accidental disturbing causes. We do not 

 consider this point as satisfactorily settled. It is to be noted that the 

 M.H.D. of the IB is practically the same, with the two complements 

 used, provided that the natural IB of the guinea-pig's serum for the 

 ox's corpuscles be first removed. In other cases recently investigated 

 by us where this does not hold, a distinct difference between com- 

 bining relationships and haemolytic action has been brought out. 



Conclusion. 



It is seen (1) that the firmness of combination is much greater in the 

 case of C with E + IB molecules than of C with anti-C, and (2) that 

 differences in the combining affinities of the C's of two animals brought 

 out when tested with anti-C may not obtain when tested with B + IB. 

 The latter result involves an apparent contradiction, but is probably 

 to be explained by the energy of combination in the two cases. It is 

 quite intelligible that differences shown to exist where the chemical 

 union is of a loose nature may not be detectable when the combining 

 affinity is strong. The relation of a toxin or complement to its 

 corresponding anti-substance is often compared to that of a key to a 

 lock, but the above results show that this is not accurate, as the com- 

 parison would involve that two different keys can be forced into the 

 same lock. Complements which may be shown by their behaviour to 

 anti-complements to be different, may still act as if they were similar 

 m the processes of haemolysis and, probably, of bacteriolysis. 



II. On Complementoids and their Combination. 



Ehrlich has pointed out the analogy which exists between comple- 

 ments and toxins of various kinds. Looked at from the point of view 



