THE PRECIPITIN REACTION. IO3 



through the use of the precipitin reaction. For the results 

 obtained by it can be expressed in a quantitative manner, and 

 are therefore entirely objective and independent of the inter- 

 pretation of the observer. The use of the precipitin reaction 

 must, however, be carefully controlled. Though this reaction 

 has been shown to be very sensitive, nevertheless it is subject 

 to a fifty to one hundred per cent, possible error as used in these 

 experiments. Furthermore it is incapable of distinguishing 

 between closely related animals. It is affected by the H-ion 

 and salt concentrations, but most important of all it must be 

 standardized through the use of secondary antigen solutions of known 

 protein content. With these precautions the results obtained 

 through its use by everyone will become strictly comparable and 

 of considerable value. 



From the data on relationships obtained with this limited 

 series of animals it is felt that Nuttall's statement is essentially 

 true. For the degree of reaction between an antiserum and 

 various animal bloods has in general paralleled the relationships 

 among those animals. Thus nearly always the reaction of a 

 given antiserum is greater with animals belonging to the same 

 order than with members of other orders. The only exception 

 to this has been the horse serum (Perissodactyla) which some- 

 times was reacted upon more strongly or equally strongly by 

 an antipig or an antibeef or an antisheep serum than were some 

 other sera belonging to the order Artiodactyla. The relationship 

 between Artiodactyla and Perissodactyla has always been con- 

 sidered to be close, the two usually being included in one order, 

 the Ungulata. The precipitin reaction reveals this close relation- 

 ship. The results on horse serum seemed irregular, however, 

 and not too much emphasis is to be put upon them. 



When essentially the same series of reactions may be obtained 

 from various antisera of the same kind, whether they be produced 

 in fowls or rabbits, the probability that these reactions correspond 

 to some reality in the relationship of the animals concerned 

 becomes very great. And when, furthermore, the reciprocal 

 values obtained with two or more different antisera check within 

 the range of error, as happened in the majority of these titrations, 

 the probable correctness of this conclusion becomes still greater. 



