83 
immunity unit which is contained in 1 c. c. of the test serum men- 
tioned in paragrapli 2, diluted 17 times. This quantity (1 c. c.) of 
serum is mixed with increasing amounts of the diphtheria poison, and 
the mixture inoculated into a series of animals, in order to determine 
with the greatest possible accuracy the least amount sufficient to cause 
the death of the animals in the first four days. This quantity of the 
diphtheria poison is then taken as the quantity to be used for the 
present test dose. Varying quantities of toxin are again added to the 
same quantity of serum (viz, 1 unit) in order to determine a second 
limit, namely, that quantity of diphtheria poison which must be added 
to the serum in order that the poison will be exactly neutralized as 
indicated by inoculating the mixture into guinea pigs. 
4. The estimation of the strength of a diphtheria antitoxin is accom- 
plished by means of a test dose of the poison (see paragraph 3) in this 
way: The particular test dose — for example, 0.355 c. c. — which is the 
present test dose of the toxin determined at the Institut, is added to 
or mixed with 4 c. c. of a mixture containing one unit of the above- 
mentioned serum. 
As the test dose of the poison is determined against 1 c. c. of a nor- 
mal serum, or against 4 c. c. of a one-fourth normal serum, conse- 
quently in testing a serum of (supposed) x strength the test dose of 
poison is added to 4 c. c of a dilution of that serum, or against 
4 c. c. of a 4^0 dilution of a serum of (supposed) one hundred times 
normal strength. 
5. The mixture obtained is injected subcutaneously into guinea pigs 
weighing 250 to 280 grams. If the animals tested l)y two members of 
the Institut die within the first four days, then that particular serum 
does not contain the strength claimed for it. If the animals die within 
the fifth or sixth day, then the serum is exactly on the limits of admis- 
sibility, and in order to prevent the necessity of soon recalling the 
serum from the market it is recommended to the manufacturer to add 
5 to 10 per cent. Indurations that appear in the test animals are not 
taken into consideration in this test. 
An autopsy is performed upon animals that die, to determine whether 
complicating diseases are present, such as tuberculosis, pseudotuber- 
culosis, or pneumonia, which may have rendered the animals more 
susceptible. 
6. Either the liquid or solid poison may be used for testing as long 
as the two limits defined in paragraph 3 have been properly determined 
and the difference between the two limits does not exceed 15 X MLD. 
Liquid poisons preserved with toluol must be used only under the fol- 
lowing conditions: 
(1) If after prolonged examination it is known that the stability of 
the test dose is constant and (2) when the test dose does not exceed 
