CASTLE, — METAMERISM OF THE HIRDDINEA. 287 



Whitman ('92) has shown that in the genus Glossiphonia (Clepsiue) 

 each distinct ganglion gives oti' three pairs of nerves, wliich are dis- 

 tributed respectively to the ring in which the ganglion typically lies 

 (sensory ring), the ring before it, and the ring behind it (Fig. 1, VIII.).* 



If we take neuromeres as criteria of somites, the three rings thus 

 innervated from a common souree should be regarded as constituting a 

 somite. According to the commonly accepted view, however, the last 

 two of these rings with the nest following ring of the body, which is 

 innervated from a different ganglion, represent a somite. See Figure 1, 



VI r. 



The only other view possible would be to regard the somite as com- 

 posed of a sensory ring and the two rings which precede it, but I am not 

 aware that such a view has ever been suggested, and I am unacquainted 

 with any facts which could be presented in its support. Accordingly 

 discussion may be limited to the two views before stated, which are 

 graphically presented at the right and left margins respectively of 

 Figure 1. 



a. Somite Reduction in Glossiphonia. 



What first suggested to me the possibility that neuromere limits may 

 coincide with somite limits, was a study of the method of somite reduc- 

 tion at the ends of the body of Glossiphonia. See Figure 1, I. -IV. ; 

 Figure 2, XXV.-XVII. In the regions indicated the metameric sense 

 organs occur, not as in the middle of the body on every third ring, but 

 on every second ring, or even on successive rings. This fact shows that 

 in those particular regions the somite contains but two rings, or, in 

 extreme cases, only one. If we regard such somites as having had 

 originally each its full quota of three rings, we must suppose that one or 

 two of the three primitive rings have subsequently disappeared. This 

 disappearance might be explained as due either to a complete suppression 

 of a ring or to its fusion with an adjacent ring. A comparative study of 

 the abbreviated somites in different species of Glossiphonia shows that 

 the latter process is invariably the first step in the abbreviation of a 

 somite. An examination of Figures 1 and 2 will make this clear. In 

 the case of somite IV. (Fig. 1) a fusion has taken jjlace between the 

 sensory ring and the ring which precedes it. A similar union has 

 occurred in somites III. (Fig. 1) and XXV. (Fig. 2). In the case of 



* This leaves out of account a branch {d, Fig. 1) which arises from the posterior 

 of the three nerves and is distributed to sense organs on the dorsal surface of all 

 three rings innervated by the ganglion. 



