TOXIN AND ANTITOXIN: METHODS OF THEIR STUDY. 557 



increasing toxicity on the fractional addition of toxin can be demon, 

 strated must at once be excluded from any mathematical analysis 

 based on a formula of equilibrium derived from the law of mass action. 

 In all of the cases 1 examined for the purpose (diphtheria poison, 

 tetanolysin, ricin, staphylolysin, arachnolysin, rennin,and precipitin), 

 this method has shown that the conception of Arrhenius and Madsen 

 is entirely inapplicable. 



The phenomena observed, however, are very readily explained 

 by the assumption of a plurality of combining groups in the poison 

 solution. Thus if to an excess of antitoxin a small quantity of 

 poison is added, as is done in the fractioning experiment, the result 

 would be that even the constituents possessing a feeble affinity and 

 which are of no consequence so far as any toxic action is concerned, 

 would be bound by the antitoxin. When then the second portion 

 of poison is added, it will be impossible for the toxin molecules, al- 

 though possessing a higher affinity, to crowd the previously bound 

 constituents .out of their combination with the antitoxin. The result 

 is that a certain portion of toxin, which would have been neutralized 

 by the antitoxin if all the poison had been mixed with the antitoxin 

 at once, now remains free. That is to say, the fractional method of 

 adding the poison has resulted in an increased toxicity, the L-j- dose 

 being reached with a smaller amount of poison. Furthermore it is 

 possible, by means of suitable technique, to cause a reduction of the 

 L dose, from which it follows that the L serum mixture contains 

 free non-toxic constituents capable of binding antitoxin, and that 

 these must possess still less affinity than the toxon. These are the 

 so-called "epitoxonoids" of von Dungern. The discovery of the 

 epitoxonoids also offers an easy explanation of the fact that it is 

 possible to immunize with mixtures of toxin and antitoxin which are 

 physiologically neutral. 



All this shows that a complete reversibility, even of the individual 



ing side reactions which do not interfere with the main reaction when one works 

 rapidly. For, as was pointed out by von Dungern and Sachs, the increased 

 toxicity is already demonstrable at a time when the union of toxin and antitoxin 

 is not yet ended. The hypothetical "side reaction" would therefore proceed 

 just as quickly as the main neutralizing reaction. 



1 The single exception met with, namely cobra venom, only proves the rule; 

 for cobra venom (we are dealing with the hsemolytic portion which is activated 

 by lecithin) is a simple substance w r ith a strong affinity for the antitoxin, as can 

 be seen from the course of the neutralization curve, which is a straight line. 



