THE PRIMARY SEGMENTATION. 127 



In choosing an advantageous animal for observation, we should 

 take a vertebrate that has undergone relatively few modifica- 

 tions. Of course, any living animal is very far removed from 

 the ancestral type, but we should expect to find the closest 

 approach to ancestral conditions in the simplest ones. The 

 sharks present many generalized features and we may as well 

 begin with them. 



It is to be understood that I am responsible for the observa- 

 tions that follow, as they have not, as yet, been substantiated 

 by any other observer. The careful examination of an elasmo- 

 branch embryo in an early stage of development shows the 

 existence of segmental folds that extend from the extreme 



J5? 



FIG. i. Very young embryo of Acanthias showing primitive segments. 



anterior end to the posterior limit of the embryo. Fig. I 

 represents such an embryo. The specimen from which the 

 sketch was made had attained a length of i.i mm. The 

 axial part of the embryo is established ; its anterior end is 

 rounded and slightly broader than the rest of the embryo. 

 There are eight pairs of segments in the axial embryo, and 

 they extend beyond into the blastodermic rim. If these seg- 

 mental folds occurred only in isolated cases, or in a single 

 embryonic stage, we should attach no especial significance to 

 them, but they are present in all normal specimens, and their 

 history shows their segmental importance. Once established, 

 they may be traced onwards in unbroken continuity and finally 

 identified with the neuromeres described by Orr, McClure, and 

 others. There are three or four pairs of mesoblastic divisions 

 in this stage that occupy a limited area in the narrow part of 



