298 PROCEEDINGS OF THE AMERICAN ACADEMY. 



three rings, but oftener of four. The sensory ring, which contains a 

 res^iiratory vesicle, is always the broadest ring of the somite. The ring 

 which is usually present, but sometimes wanting, is a narrow one, which 

 has been added at one end of the somite. It is impossible to say, without 

 studying the nerve distribution, at which end it has been added. Appar- 

 ently, however, it is at the anterior end, for in those animals which I 

 have had an opportunity to examine, the new ring appears to be united 

 more closely with the ring which precedes than with that which follows 

 a sensory ring. Moreover, the ring which precedes the sensory ring is 

 usually not so broad as the one which follows it. This is an indica- 

 tion that it is the former rather than the latter which has undergone 

 division.* 



In Trachelobdella (Fig. 6, F), according to Blanchard. the somite con- 

 sists sometimes of three, sometimes of six rings, each of the three prim- 

 itive rings having, in the latter case, divided. The sensillse remain on 

 the anterior derivative of the original sensory ring. 



In Cystobranchus (Fig. 6, G) the somite consists of seven rings, two 



of which bear the respiratory vesicles and doubtless have arisen, as in 



Trachelobdella, from the division of the original sensory ring. The 



remaining five rings have probably been derived from the two primitive 



non-sensory rings, — three from one non-sensory ring, two from the 



other. One of the non-sensory rings must, accordingly, have divided 



only once, as in Trachelobdella ; the other, twice. The position of the 



genital pores would seem to indicate that the posterior non-sensory ring 



is the one which has divided a second time; but it is impossible to say 



without study of internal structures whether this is really the case or 



not. 



A typical somite of Piscicola contains, according to Apathy ('88), 



twelve annuli ; according to Blanchard ('94), fourteen. Not having 



had an opportunity to examine this genus myself, I do not venture to 



express an opinion as to the number or morphological value of the rings. 



Blanchard, however, states that the respiratory vesicle is borne on two 



rings of the somite, the anterior of which contains the nephropore. This 



would seem to indicate that the sensory ring had divided only once, and 



that the remaining twelve rings (or ten, Apathy) had arisen by repeated 



division of the two primary non-sensory rings. 



To recapitulate : — We find in the Rhynchobdellidte that, starting with 



a somite of three rings of equal width, increase in the number of annuli 



* Vaillant ('70), who was guided by tlie position of tlie "septa," also placed the 

 Bomite limits so as to make two rings precede the sensory ring and one follow it. 



