43 
The substance called anaphylactin, which was demonstrated by 
Gay and Southard, could not properly be considered as an antibody, 
as the sensitizing action of the transferred serum could very prob- 
ably have been due to a horse serum “rest,” as suggested by those 
authors. 
However, Otto a subsequently showed that if untreated guinea 
pigs be given an injection of serum of a sensitive guinea pig and 
tested for susceptibility twenty-four hours later, they will be found 
to be sensitive. This sensitizing substance, demonstrable within 
twenty-four horns, is very probably an antibody. 
In our first paper b we suggested, as a working hypothesis, that 
the first injection of the serum results in the formation of an anti- 
body which, when brought into contact with more serum at the 
second injection, produces a union or reaction, resulting in the toxic 
manifestations. 
It occurred to us that if the serum of a sensitive guinea pig be 
mixed wdth normal horse serum and its toxicity tested upon sensi- 
tive animals perhaps there might be evident an increase in the toxicity 
of the horse serum. 
In order to test this hypothesis we bled sensitive guinea pigs 
which had previously received a small amount of normal horse 
serum. This blood was defibrinated and the serum mixed with an 
equal volume of normal horse serum and allowed to remain at room 
temperature for one hour, when the toxicity of the mixture was 
tested upon sensitive guinea pigs. 
The blood serum of the sensitive guinea pigs used in each one of 
these experiments was shown by appropriate tests to contain a 
sensitizing substance for untreated animals. 
As controls we gave sensitive guinea pigs an amount of horse serum 
equal to that contained in the mixture of sensitive guinea pig blood 
and horse serum. As further controls we also gave sensitive guinea 
pigs an equal amount of a mixture of normal guinea pig serum and 
normal horse serum, which had been allowed to remain in contact for 
one hour. The details of the experiments will be seen in Table 
No. 20. 
In order to test the toxicity of such mixtures, it is necessary to 
determine the minimum lethal dose of horse serum, so that the con- 
trol animals will receive a dose sufficient to cause symptoms in most 
of them, but not death; so that if there is any increase in the toxicity 
it will be readily manifest. 
It appears from a limited amount of work, which we present in the 
following tables, that there was an increase in the toxicity of the 
“Munch. Med. Woch., vol. 54, 1907, p. 1665. 
&Hyg. Lab. Bui. No. 29, 1906. 
