STANDARDIZATION OF SERUMS. 183 



been found, for example, that following each in- 

 jection of toxin the amount of antitoxin in the 

 blood suffers a reduction, and only equals or rises 

 above the previous amount eight or ten days later. 

 This decrease is explained by assuming that the 

 toxin has, to a certain extent, united chemically 

 with the circulating antitoxin. It indicates also 

 the period at which the horse should be bled in 

 order that the greatest amount of antitoxin may 

 be obtained. It might even be dangerous to draw 

 the blood before this time had elapsed, since some 

 free toxin might still be in the circulation. 



It is noteworthy that all horses are not equally 

 good producers of antitoxin. One may yield a 

 serum of three times the value of another, al- 

 though the two have been treated identically and 

 seem to be equally immune to the toxin. 



Another most interesting fact is that, although 

 the blood of an animal may be very rich in anti- 

 toxin, he still may have a disproportionate sus- 

 ceptibility to fresh injections of the toxin. 



Many of these phenomena have not been ex- 

 plained satisfactorily. 



The necessity of standardizing antitoxins so 

 that dosage may be controlled accurately is self- Sjf 

 evident. To meet this need the antitoxic unit toxins - 

 familiar in practice was devised. 



Behring, and also Ehrlich, decided arbitrarily 

 to consider as the antitoxic unit that quantity of a 

 serum which would protect a guinea-pig from 100 

 fatal doses of the toxin. Ehrlich's original method 

 of testing a serum was to mix different quantities 

 with 10 fatal doses of the toxin and inject each 

 mixture into a guinea-pig of from 250 to 300 

 grams' weight. That quantity of the serum which 



