8 
expiration of a certain interval is again injected with horse serum 
the result will probably be fatal. The first injection of horse serum 
has sensitized the animal in such a way as to render it very suscep- 
tible to a toxic principle in horse serum. It is probable that when 
the guinea pig is injected with the first, or sensitizing, quantity of 
serum the strange proteid contained in the horse serum develops in 
the body of the guinea pig “ antibodies ” which, when brought into 
contact with more horse serum given at a second injection, produce 
either a union or a reaction, which causes the toxic action. 
A certain time is necessary to elapse between the first and second 
injections of horse serum before this toxic action is able to manifest 
itself. This 11 period of incubation” is from ten to twelve days, 
and corresponds suggestively with the period of incubation of the 
serum disease which Pirquet and Schick place at eight to thirteen days. 
Guinea pigs may be sensitized with exceedingly small quantities 
of horse serum. In most of our work we used quantities less than 
2J0 c. c. and we found in one instance that T . o o o\oo o c - c - of horse 
serum was sufficient to render a guinea pig susceptible. 
It also requires very small quantities of horse serum, when given 
in a second injection, to produce poisonous symptoms. One-tenth 
c. c. injected into the peritoneal cavity is sufficient to cause the death 
of a half -grown guinea pig. One-tenth c. c. of horse serum injected 
subcutaneously is sufficient to produce serious symptoms. The fact 
that this toxic action may be developed by such small quantities 
of serum and the fact that exceedingly small quantities are sufficient 
to produce symptoms and death upon a second injection, a priori 
places both the sensitizing and the toxic principle in the horse serum 
in the “ hap tin group” of substances in the sense used by Ehrlich. 
A still further indication that the side-chain theory in its broadest 
sense may be applicable is the further fact that immunity may be 
produced against the toxic action by multiple injections of the 
serum. 
While at first we thought that diphtheria antitoxin had some 
relation to this action, we are now able to state positively that it 
has nothing whatever to do with the poisonous action of horse serum; 
further, that diphtheria antitoxin in itself is absolutely harmless. 
The toxic action which we have studied is caused by a principle in 
normal horse serum and is entirely independent of the antitoxic 
properties of the serum. 
