424 Drs. C. J. Martin and T. Cherry 



A standardised solution of diphtheria toxin was filtered through 

 such a filter. The filtrate was found to contain diphtheria toxin. 

 This filtrate was then tested to ascertain whether it were as toxic as 

 the original solution. As will be seen from the protocols it was 

 somewhat diminished in toxic power (Protocol I). 



The antitoxin of diphtheria, as was shown by Brodie,* does not 

 pass through such a filter. When antitoxic serum is filtered through 

 gelatin, the whole of the proteids, and together with them all anti- 

 toxic virtue, are absent from the filtrate (Protocol II). As the toxin 

 is not held back by the filter, whereas the antitoxin is, one is provided 

 with a simple physical means of separating them, provided they have 

 not reacted upon one another. 



We mixed a solution of toxin containing eight fatal doses per 

 kilogram of guinea-pig in each c.c., with sufficient Behring's anti- 

 toxin to more than completely neutralise all the toxin. This mixture 

 was allowed to remain in contact at 30 C. for two hours, and then 

 filtered through the gelatin filter. Varying quantities of the 

 filtrate were injected into guinea-pigs up to nearly 4 c.c. per kilo- 

 gram of body weight, that is a quantity originally containing 32 

 fatal doses. The filtrate was quite innocent. The guinea-pigs 

 suffered no inconvenience, and gained weight while under observa- 

 tion in small cages. The injections produced no local oedema. 



If the toxin had remained unaffected beside the antitoxin there 

 was nothing to prevent it passing through the filter in virtue of its 

 relatively small molecular size. As, however, it did not do so, we 

 can only conclude that it had entered into some sort of chemical 

 relationship with the relatively large molecules of the anti-toxin 

 during their sojourn together prior to filtration. 



Having obtained results so definite, and in apparent contradiction 

 to those of the authors quoted in the beginning of this paper, we 

 next experimented with snake venom in order to repeat Calmette'sf 

 observations. 



We took a series of rabbits (Protocol V) and injected them with 

 mixtures containing one constituent of the venom of Eoplocsphalus 

 curtus and Calmette's antivenomous serum. On reference to the 

 protocols of this series of experiments it is seen that 2 c.c. of this 

 sample of serum was sufficient to counteract an amount of the poison 

 contained in '0002 gram of the dried venom. This amount killed 

 control rabbits in about eight hours (Protocol IV). 



In some of the experiments this amount of venom and serum was 

 allowed to remain in contact for fifteen minutes at the laboratory 

 temperature (21 C.) and then heated to 68 C. for ten minutes to 

 destroy the antitoxin. In Calmette's experiments the rabbits 



* Journ. of Path.,' 1897, p. 460. 

 + Cairn ette, loc. cif. 



