198 INFECTION AND IMMUNITY. 



toxin, a deuterotoxin and a tritotoxin may be rec- 

 ognized by this same partial saturation method. 

 (See Fig. 2.) For example, it may be found that 

 when a portion of the antitoxin unit, between the 

 limits of 149/200 and 125/200, is withheld, a 

 toxin is left free which is less virulent than that re- 

 maining free between the limits of 124/200 and 

 100/200 ; and from this point on the new unbound 

 toxin may be still more virulent. The first would 

 be tritotoxin, the second deuterotoxin and the 

 third prototoxin. 



The "spectrum" of a toxin changes with its age. 

 The prototoxin, and portions of the deutero- or 

 tritotoxin may disappear because of toxoid forma- 

 tion. Such changes have led to the recognition of 

 an alpha and a beta modification of the toxin. 

 The alpha modifications of all three toxins readily 

 become toxoids. Only the beta modification of the 

 deuterotoxin remains constant. The toxon also 

 remains relatively intact (Figs. 2, 3 and 4). 



This very complicated method of investigation 

 was also undertaken by Madsen in the study of 

 tetanus toxin, for which a somewhat similar 

 "spectrum" was constructed. 



Such spectra have not been worked out in de- 

 tail for some of the vegetable toxins, as ricin and 

 abrin, but it is known that they form toxoids. 



Some of the toxins of snake venom differ from 

 the bacterial toxins in structure (pages 428-431). 

 The Forma- The idea was originally advanced that antitoxin 

 Antitoxin' is transformed toxin, a change in the latter hav- 

 ing been effected through some action of the 

 tissues. In that case,, the amount of antitoxin pro- 

 duced should be roughly equivalent to the amount 

 of toxin injected. This, however, was found not 



