102 ALAX ARTHUR BOYDEN. 



of animals we have failed to reach agreement as to what consti- 

 tutes a species or to learn how to express quantitatively the 

 relations of one species or genus or family to another. For 

 classification has been based chiefly on adult morphology, the 

 tendency being to throw aside the evidence from developmental 

 morphology as being more liable to error. However this may be, 

 morphology itself is subject to certain limitations. Chief among 

 these are two (i) an error due to convergence which it is difficult 

 to estimate with certainty, (2) the failure to give a quantitative 

 expression to morphological features in general and hence the 

 necessity of depending much on interpretation as to what various 

 structures may mean in descent. Interpretations differ with 

 interpreters and hence it is that there is endless difference of 

 opinion as to the relationships of certain groups of animals 

 necessitating countless "revisions" of them. The discovery of 

 an objective, quantitative test which could be applied to animals 

 and through which some general agreement as to their relation- 

 ships might be attained would be a great help to the zoologist. 



Attempts to study other than gross morphological structures of 



animals for the purpose of understanding their relationships have 



already been made. The study of the distribution of respiratory 



pigments and related substances is an instance in point. While 



it has been found that hemoglobin is rather erratically distributed 



among Invertebrates, the gaps may partly be rilled with animals 



possessing related substances. The study of the crystallography 



of the hemoglobins by Reichert and Brown (1909) is another 



The results obtained by these authors confirmed in 



general the existing classification. Chemical analyses of various 



tissues ha\c aKo been made but such methods fail for the most 



part to show stereochemic differences between substances of 



essentially the same elementary constitutions. It seems to be 



true, however, that it is generally such stereochemic differences 



\\hicli are most important in the study of 



\\- mii-i have a method therefore which may 



in. imitative expression to these differ- 



"ncl: 3 available in the precipitin reaction. 



ilu-n tin- niM-i important contribution to the study of 



i'lr from morphology, can be made 



