114 BACTERIOLOGY. 
tected from access of air, the more resistant antitoxins 
may be preserved sometimes for a year or two with 
practically no deterioration in strength. At other 
times, however, from unknown causes, they are gradu- 
ally destroyed, so that there may be a loss of about 10 
per cent. per month. A serum requires, therefore, to be 
tested every few months if we wish to be assured of its 
strength in antitoxin. Preservatives, such as carbolic 
acid, trikresol, camphor, etc., alter antitoxins only very 
slightly when in dilute solution, but in strong solution 
they partially destroy them. Heat up to 62° C. does 
not injure them greatly, but higher temperatures alter 
them. In animals injected with diphtheria toxin 
Atkinson has found that there is with the increase in 
antitoxin an almost proportional increase in globulin.? 
He also found that the antitoxins behave like globulins 
with the various reagents, being completely precipi- 
tated by magnesium sulphate. NaCl, when added to 
saturation to globulin solutions holding antitoxin, parti- 
ally precipitates the globulin and the antitoxin. When 
raised to 72° C. a series of precipitations are obtained 
which contain at least the greater part of the antitoxin. 
Whether this indicates that antitoxin is a form of 
globulin or merely that it is similarly affected by many 
reagents, and that the toxin in some way stimulates the 
development of both, it is as yet impossible to say. 
Although in a very rough way, the same animal pro- 
duces antitoxin in direct proportion to the amount of 
toxin injected so long as its condition remains good, yet 
different animals of the same species give very varying 
amounts from the same injections, some not giving one- 
1 This work, carried out in the Research Laboratory of the Department of 
Health of New York City, will appear in the Journ. of Exp. Med. in 1900. 
