5 IS STUDIKS IX IMMUNITY. 



the jiowcr of adsorption varies with each adsorbing substance and 

 each adsorbed Bubstance. It depends, certainly, on physical quali- 

 ties but also on the chemical nature of the substances concerned, 

 and would appear to us more variable still if we could measure by 

 delieate methods in each instance. Certain colors act on substances 

 like paper more rapidly than others, and the subsequent decoloration 

 of the paper in a large volume of water is sometimes easy and some- 

 times difficult. In the adsorption of albuminous substances by 

 inorganic colloids all degrees of reversibility are found. In serum 

 (reactions a similar irregularity occurs. Certain antibodies act very 

 rapidly and others require more time to unite with their antigens 

 and the complexes obtained are dissolved with very unequal facility. 

 As is evident, particularly from the researches of Morgenroth, that 

 the neutralizing of diphtheria toxin by antitoxin is much slower than 

 was at first believed. Certain complexes, such as those formed by 

 agglutinins or sensitizers with bacteria or corpuscles, or by certain 

 toxins with their antitoxins, are, if not completely, at least dis- 

 tinctly, reversible. In this relation I may recall the work of Hahn 

 and Tromsdorff, Landsteiner, Morgenroth, Muir, Madsen, Otto and 

 Sachs. But it should be noted that, in general, the complexes of 

 antibody-antigen, especially when they have been standing for some 

 time, manifest a very imperfect reversibility, distinctly less than 

 Arrhenius and Madsen's thesis would demand. In the same way, 

 in general, adsorption phenomena give rise to incompletely reversible 

 complexes. As we know the study of the neutralization of toxins 

 by antitoxins has brought out the fact, particularly owing to the 

 important researches of Ehrlich, that it is frequently impossible to 

 prepare mixtures which are exactly neutral. It is likewise known, 

 so far as diphtheria toxins and antitoxins are concerned, that Ehrlich 

 explains this fact by supposing that the toxin contains in reality 

 several poisons. The same phenomenon, however, occurs in in- 

 stances where there is no reason for supposing that such a com- 

 plexity in the toxic fluid exists. In such instances there is a priori 

 no reason for agreeing with the thesis of Arrhenius and Madsen in 

 accordance with which the reaction is incomplete and leaves a cer- 

 tain amount of the toxin free, or with my own theory in accordance 

 with which the molecules of toxin and antitoxin may unite in vari- 

 able proportions according to the respective quantities of the two 



