COLLOIDAL THEORY OF ANTIBODIES 329 



two, this mixture did not regain toxicity. Several lethal doses 

 must first be added, just as it is necessary to add several lethal 

 doses of diphtheria toxin to the L dose. 



A difficulty arises from the fact that both toxin and antitoxin 

 pass in the same direction in an electric current. These experi- 

 ments are naturally difficult to carry out, since the effect of 

 electrolysis must be eliminated. When this is done, according to 

 Field and Teague, both substances travel towards the cathode, and 

 this whether the solution is acid or alkaline. But we must 

 remember that when toxin and antitoxin interact they always do 

 so in a very complex fluid, containing many other substances, 

 both colloids and electrolytes, and until we can determine the 

 electric charge of these substances in a pure form, these experi- 

 ments can hardly be considered to outweigh the remarkable 

 analogies described above. It is highly probable, judging from 

 analogy with the other antibodies, that free ions are essential to 

 the neutralization of toxin by antitoxin, and the researches of 

 Girard-Mangin and Henri show us how complex the conditions 

 may become. 



We have seen that it is important to know whether the 

 compounds of antibody and antigen are dissociable, and that the 

 latter view explains some difficult phenomena seen in immunity. 

 On Ehrlich's theory the combination between the two is a strong 

 one, and once it has been allowed to take place, the two substances 

 are not dissociated into their components, whereas on Arrhenius's 

 and Madsen's theory the combination is an unstable one, and dis- 

 sociation constantly occurs. The experimental proof in favour of 

 the occurrence of this phenomenon appears on the whole satisfac- 

 tory. Red corpuscles saturated with immune body mixed with 

 normal corpuscles will part with some of their antibody, so that 

 the latter become sensitized. A mixture of tetanus toxin and 

 antitoxin, which causes no symptoms when injected into the leg of 

 a guinea-pig, causes tetanus when adrenalin has previously been 

 injected locally (explicable on the fact that toxin is absorbed by 

 the nerves and antitoxin by the vessels, which in the second case 

 are constricted), and other examples might be quoted. It appears, 

 however, that this irreversibility only occurs if the two substances 

 are not kept in contact for a sufficiently long time, as was pointed 

 out by Martin and Cherry in their filtration experiments very 

 early in the history of the chemistry of the action of antitoxin. 

 Thus in the latter example, if the tetanus toxin and antitoxin be 



