74 
been standardized by the Bureau of Standards in which the total 
amount of the serum may be accurately read to the fractional part of 
a cubic centimeter. 
From 0.5 to 1 c. c. of each lot of serum is planted into a large amount 
of bouillon in order to dilute the preservative. This is incubated at 
body temperature and an^^ growth that may appear is determined b}^ 
the usual bacteriological methods. 
The number of units that the serum contains is then estimated by 
the following procedure: 
If a syringe contains 4 c. c. of serum and is labeled 1,000 units it 
must of course have at least 250 units to each cubic centimeter. We 
would then test such a sample upon four or live guinea pigs to see 
whether its unit strength is above or below the value claimed for it. 
If the strength of the serum is 10 per cent below that claimed, a special 
report is made to the Surgeon-General to this etfect, the manufacturer 
is notified that all the serum bearing this laboratory number must be 
withdrawn from the market, and the matter is investigated. 
The tests to determine the strength of commercial serums sold upon 
the open market are made against a toxine that has been very carefull}^ 
tested to determine its exact L+ dose. Controls at the same time are 
always made against both Ehrlich’s normal serum and the official 
standard serum made in this laboratory. In case of doubt or contra- 
dictory results the work is always repeated, making the second tests 
against the L+ doses of two toxines. 
As weak serums are of doubtful therapeutic value, we do not favor 
the sale of diphtheria antitoxic serum containing less than 250 units 
per c. c. If a serum bought on the market contains less than 200 
units per c. c., the fact is reported to the Surgeon-General, who 
requests its withdrawal from sale and a discontinuance of the practice. 
To determine that an excessive amount of preservative belonging to 
the phenol group has not been added, the following tests, modified 
from similar tests suggested by Ehrlich, are employed: 
It has been found that 0.5 c. c. of a 0.5 per cent solution of phenol 
will kill a medium-sized mouse if injected subcutaneously. The quan- 
tity of phenol contained in 0.5 c. c. of such a solution represents 0.0025 
gram. It requires 1 c. c. of a 0.5 per cent solution of tricresol to pro- 
duce the same result. One c. c. of a 0.5 per cent solution contains 
0.005 c. c. of tricresol. The above facts have been experimentally 
confirmed in this laboratory. 
As phenol is not used as a preservative in this country for diph- 
theria antitoxic serums, preference being given by manufacturers to 
tricresol, we inoculate mice subcutaneously with 1 c. c. of their serum. 
Mice so inoculated sometimes show trembling and other manifesta- 
tions of tricresol poisoning, which shows that the serums contain very 
