38 
pigs within four days. e not inf requenth" obtain toxines, the MLD 
of which is 0.005 c. c. However, such favorable results are not ^ 
always obtained. Sometimes for several weeks in succession our 
toxines have failed to kill guinea pigs in doses as large as 0.01 c. c. 
It must be evident to anyone who has had practical experience with 
the manufacture of diphtheria antitoxin that there are factors con- 
cerned in the production of a strong poison that are not at all under- 
Fig. 5. — Method of filtering the culture. Flask for storing the toxine in bulk with Maasen mouth 
' piece. 
stood. It is sometimes discouraging to those who take especial pains 
to comply with all the requirements believed to favor the production 
of strong toxines to find that other laboratories who use simpler 
methods obtain, on the average, poisons of equal and sometimes supe- 
rior toxicity. 
Roux and Yersin« noticed that the toxicity of a liquid increased more 
rapidly and more regularly Avhen the culture is grown in the presence 
«Mace, E, : Traite pratique de bacteriologie. pp. 589-590. 
