Jan. 8,1917 
Immunity Studies on Anthrax Serum 
47 
beakers, and heated as before to coagulate the protein. Only a small 
quantity of acetic acid is necessary for flocculation—usually 0.5 to 1 
c. c. of the N!$ acid was sufficient. The precipitated protein is then 
filtered on weighed papers, washed till free, or almost free, from sulphate, 
then washed with small amounts of alcohol and ether, dried, and weighed 
as in the determination of total coagulable protein. 
If desired, the globulin precipitate may be dissolved in 40 c. c. of 
water and again precipitated by the addition of 50 c. c. of saturated 
ammonium sulphate. This may free the precipitate of traces of albumin, 
but the loss of globulin at the same time probably makes this an un¬ 
necessary step. 
Evaporation from the free surfaces of the fluids in the tubes during a 
40 minutes’ run in the centrifuge was found to be negligible, amounting 
to less than 1 or 2 c. c. 
Good duplicates are easily obtainable. In 23 determinations the 
duplicates differed from 1 to 16 mgm., with an average of 7 mgm. This 
does not include a few determinations that were repeated because the 
duplicates differed enough to indicate error. Much depends upon the 
condition of the centrifuge. This must be a high-speed, smooth-running 
apparatus, which slows down smoothly. Practically the same results 
are obtained when the same serum is used for two globulin determina¬ 
tions about one month apart. The serums were preserved with 0.5 
per cent chloroform (likewise the globulin preparations) and kept in a 
refrigerator. 
The object of precipitating one volume of serum in a final dilution of 
10 volumes of one-half saturated ammonium-sulphate solution is to pre¬ 
vent the contamination of the precipitate with albumin, which is said to 
be absorbed. This is the reason why the precipitate is dissolved and 
reprecipitated by some workers. While this procedure may be advisable 
for certain analytic purposes, it is objectionable when such results are to 
be used in connection with a study of antitoxin or similar products 
obtained by precipitating one volume of serum in a final dilution of three 
volumes of one-half ammonium-sulphate solution. This is one of the 
reasons why the analytic method has been further modified and improved 
so that the precipitation of globulin in the analyses and in the separation 
of large quantities of globulin for therapeutic use are both accomplished 
under the same conditions. 
The filter papers used were S. & S. 589 “white ribbon,” 15 cm. These 
were placed in weighing bottles, 50 by 40 cm., dried for six hours in an 
electrically heated air oven at ioo° C., and weighed. They were then 
dried a second time for two hours and weighed to be certain that the 
weight was “constant.” Almost invariably the second weighings differed 
from the first by 2 or 3 mgm. Drying the papers to an absolutely con¬ 
stant weight seldom occurred. Two empty weighing bottles were dried 
along with the others. Their weight, which was taken several times, 
varied only a fraction of a milligram. All weighings were to the nearest 
