522 STUDIES IX IMMUNITY. 



once, on the contrary, the antitoxin is spread over all the toxic 

 molecules and a complex is obtained which contains an even pro- 

 portion of the antidote, and which, consequently, is not as fatal as 

 even a small dose of free toxin. 



It is an extraordinary thing that, so far as precipitins and its pre- 

 cipitable antigens are concerned, this interpretation is not contested, 

 even by such writers as von Dungern, who are, in general, favorable 

 to Ehrlich's ideas. As Halban and Landsteiner, Eisenberg, von 

 Dungern, and Miiller, in particular, have seen, if we mix a precip- 

 itable serum "A" with its appropriate precipitating serum "B," an 

 excess of the precipitable substance inhibits the occurrence of a 

 precipitate. For example, it is found that in a mixture of one 

 volume of serum "A" with one volume of serum "B," a cloudiness 

 appears, but a mixture composed of two volumes of "A" and one 

 volume of "B," remains translucent. The almost obvious explana- 

 tion is that the molecules of "A" are not susceptible to precipita- 

 tion unless they have fixed a sufficient number of molecules of 

 "B." Such, however, is the condition in the first mixture. In the 

 second mixture the molecules of "B" are shared among all the 

 molecules of "A" which are in too great a number to permit their 

 perfect separation and consequently they are not precipitated. 

 In accordance, then, with the proportions of constituent parts 

 different complexes may be obtained, and this fact comes out 

 quite clearly in the experiment of the addition in a single or in 

 divided doses. If we add one volume of "B" to two volumes of 

 "A," the mixture remains limpid, even after a considerable time. 

 If, on the other hand, we add one volume of "B" to one volume of 

 "A," a precipitate immediately appears. When the precipitate is 

 completely formed we may add another volume of "A" so as to 

 make the mixture correspond to the preceding one. The precipi- 

 tate does not disappear, but remains indefinitely intact and so we 

 have obtained two mixtures composed of the same sera in the same 

 proportions but presenting entirely different aspects. 



Does not this experiment offer most satisfactory analogies with 

 the one on coloring filter paper with a dye to which I have just 

 referred? The objection may be raised that, although true as far 

 as precipitins are concerned, it may not be applicable to antitoxins. 

 Such an objection, however, would be to deny any character of 



