362 BACTERIOLOGY. 
original strength for at least two months; after that it 
canbe used by allowing for a maximum deterioration 
of 10 per cent. for each month. The antitoxin in old 
serum is just the same as in that freshly-bottled, only 
there is less of it. 
The nature of diphtheria antitoxin has until recently 
been known almost wholly from its physiological prop- 
erties. Recently experiments have seemed to show that 
it was either closely bound to the globulins or was itself 
a globulin. Mr. J. P. Atkinson, assistant chemist in 
the laboratory, has kindly permitted me to state the 
results of his investigations, which will soon appear in 
the Journal of Experimental Medicine. He found that 
antitoxicand normal horse-seruin react similarly toward 
MgSO,, in that the globulin is precipitated completely 
- from the other constituents of the serum. In the case 
of antitoxic serum the globulin precipitate carries with 
it all of the antitoxic power of the serum, leaving the 
filtrate without any neutralizing power against the diph- 
theria toxin. When watery solutions of this globulin 
are saturated with NaCl a precipitate occurs. When 
the solution is heated a series of further precipitates take 
place, as follows: Cloudiness appears at 40°, 49°, 57°, 
and 67° C.; complete precipitate occurs at 45°, 54°, 
62°, and 72° C. Lach of these precipitates has anti- 
toxic properties, and the total quantities contain all the 
original antitoxin except some 5 per cent., which is evi- 
dently destroyed by the higher temperatures required 
for the last two precipitates. After the last precipitate 
the solution is free of globulin and also of all antitoxic 
properties. 
A further fact developed by Atkinson is that the 
globulins increase markedly in the serum of horses as 
