104 



ALAN ARTHUR BOYDEN. 



Much criticism of Xuttall's work has appeared in recent years 

 in Europe. Statements have been made that the precipitin 

 reaction did not follow at all the relationships of the proteins 

 tested. The results given here do not support such views. 

 Never has any heterologous reaction in this entire series exceeded 

 the homologous reaction of the same antiserum, as Friedberger 

 and Collier (1919) or Friedberger and Meissner (1923) claim. 

 Perhaps the fault lies with their technique for they have ap- 

 parently used neither buffered salt solutions nor standard antigens 

 of known strength. Furthermore the dilutions which they made 

 of antigen were only i-ioo; 1-200; 1-1,000; 1-10,000; and 

 1-20,000. The possible error in their readings must therefore 

 be 1,000 per cent, in most cases. If such crude methods had 

 been used in these tests no doubt the results obtained would 

 correspond to theirs but the value of the precipitin reaction in 

 the study of relationships would have been entirely overlooked. 

 A better technique must be used if results of value are to be 

 obtained. It is not that poor results cannot be obtained with 

 this reaction, the important thing is that good results can be 

 obtained and poor results recognized as such and discarded from 

 further consideration. 



The fact that the values obtained in this study of a limited 

 series of animals correspond closely with the results of Nuttall's 

 quantitative tests shows that the precipitin reaction can give 

 essential uniformity of results. Having thus proven its value in 

 study of Vertebrates where relationships are relatively 

 certain, it may now be applied to the study of animals of more 

 uncertain affinities. The Invertebrates offer a large and im- 

 portant field for such research. A few results have already been 

 Xuttall in the study of Crustacea. Perhaps also 

 his demonstration of the affinity of the King Crab (Limulus) to 

 spiders is to be taken as an indication that the further application 

 of the precipitin reaction to the study of relationships of In- 

 vertebrates will eventually yield results of some value to the 

 phylogenist. 



CONCLUSIONS. 



i. The .possible error in reading the ring test in this study 

 has been 50-100 per cent. 



