296 PROCEEDINGS OF THE AxMERICAN ACADEMY. 



from the sensory ring, 3. In somite XXVII. a complete union has 

 taken place between rings 1—3 ; rings 4 and 5 are probably repre- 

 sented in the broad, curved part of the body which bears distally the 

 acetabulum. This part is imperfectly divided by transverse creases 

 (shown too distinctly in the figure), but is sharply marked ofE from the 

 preceding ring. 



To recapitulate : — We find in the case of Nephelis the same rules gov- 

 erning the process of somite reduction as in the case of Glossiphonia. 



(1). Rings fuse into groups innervated typically from a common 

 ganglion. P^xamples : Figure 5, right half, somites XXV.-XXVII. ; 

 Figure 4, somites III. and IV. The union, at the margins of the body, 

 of somite I. with the anterior ring of somite II. presents an apparent 

 exception, but here we have to do with an extreme case, in which all the 

 rings of one somite (I.) have fused with one another, and in addition a 

 fusion between successive somites (I. and II.) is foreshadowed. 



(2). The first rings to disappear are those at the ends of the somite, 

 the anterior end being usually the one which is first affected. Examples: 

 in Figure 4, somite VII. (only the most anterior ring, 1, wanting) ; so- 

 mite VI. (a ring wanting at either end of the somite) ; somite V. (two 

 rings wanting at the anterior end of the somite, only one wanting at 

 the posterior end) ; in Figure 5, somite XXVI. (rings 1 and 2 fused 

 together and partially united with ring 3, rings 4 and 5 united) ; somite 

 XXVII. (rings 1-3 united, rings 4 and 5 united). 



In somite XXV. alone we have a case where abbreviation affects the 

 posterior part of the somite sooner than the anterior part. But it will 

 be observed that the middle of somite XXV. marks the boundary 

 between the abbreviated and unabbreviated portions of the posterior 

 half of the body. This may explain why in this single instance rings 4 

 and 5 are reduced, while rings 1 and 2 are unaffected. 



From the foregoing fads it appears that in the case of Nephelis as well 

 as of Glossiphonia there are natural somite limits, which coincide with 

 the limits of the neuromeres. 



IV. Somite Growth. 



Having now examined with some care the process of somite reduction 

 in the leech, it may be instructive to study also the reverse process, 

 namely that of increase in the number of rings in a somite. This can 

 best be done by an examination of typical somites from various genera 

 of the two commonly recognized families of leeches. 



