56 THE AMERICAN MONTHLY [Feb. 



in the incubator and kept at a constant temperature of 

 37 degrees C. until the bacilli have become very numer- 

 ous, and have secreted enormous amounts of active and 

 powerful toxin in the bouillon. 



When this has taken place a microscopical examination 

 is made to see that bacilli other than the Klebs-Loeffler 

 are not present, and the diphtheria toxin thus contami- 

 nated. If uncontaminated 1 per cent of trikresol is added 

 to prevent contamination and to destroy the bacillus 

 diphthericB. The bouillon, or, as we now term it, diphthe- 

 ria toxin, is filtered through a modified Chamberland 

 filter to separate from it the dead bodies of the diphthe- 

 ria bacilli. No bacilli are therefore injected into the 

 animals to be immunized, and they are not given diphthe- 

 ria, but only the toxin secreted by the bacilli. 



DETERMINING THE TOXICITY OF THE TOXIN. 



The toxicity of the toxin is determined by its injection 

 into guinea pigs. To be of the desired strength, 0.01 to 

 0.1 ccm. should produce death of the control animal in 

 from 24 to 36 hours. 



For the preparation of diphtheria antitoxin any ani- 

 mal may be selected, but horses are preferred, inasmuch 

 as they are more easily operated upon, and because they 

 furnish excellent serum in liberal amounts. Our experi- 

 ence as to the type of horses selected, particularly in the 

 earlier observations, have been valuable, the majority 

 being of unusually high quality, a number showing trace 

 of fine breeding; such horses, however, are not suited for 

 immunization. The finely bred horse being sensitive, 

 frets at his inactivity (for no work is performed by the 

 animal while being immunized, only a suflBcient amount 

 of exercise being given to maintain good health), neither 

 does he take kindly to the injection of the toxin or the 

 subsequent bleeding operations. The preference is given 

 to large, compactly built animals, of dark color, 16 to IS 



