800 PROCEEDINGS OF THE AMERICAN ACADEMY. 



place ^rs< and most often in those rings which stand at the limits of the 

 somite (rings 1 and 5, Fig. 7). 



If one marks off the somite limits in any other way than that which 

 I have followed, regarding the sensory ring either as the first or as the 

 last ring of the somite, the pi'ocess of ring multiplication becomes unin- 

 telligible, taking place now in the middle, now at one end of a somite. 

 The reader can test this for himself by reconstructing the diagrams given 

 in Figures 6 and 7 so as to make the stippled ring come at one end or 

 the other of the somite. 



From the foregoing discussion the conclusion seems warranted that 

 neuromeric groups of rings, that is, rings innervated typically fi'om the 

 same ganglion, are natural groups behaving as units both in the process 

 of abbreviation and in that of elongation of jiarticular body regions; in 

 other words, that they are the true morphological units or somites. 



V. Primitive Condition of the Leech Somite. 



We have seen that the process of ring multiplication is very general 

 among the leeches, and that it takes place in quite an orderly manner, 

 new annuli being formed, in the great majority of cases, at the ends of 

 the somite. We have seen, further, that the complicated forms of somite 

 found in the Gnathobdellidaj and some of the Rhynchobdellida3 are all 

 derivable from a primitive three-ringed type of somite, like that of Glos- 

 siphonia. In view of these facts, does it not appear probable that the 

 three-ringed type of somite itself has been derived from a simpler primi- 

 tive condition? I am stronjjiy inclined to think so. 



As to what this simpler condition was, we perhaps may get an idea 

 from an examination of the abbreviated somites of Glossiphonia (Fig. 6, 

 B, A). For we have found that multiplication of annuli is the reverse 

 process of somite reduction, both alike affecting the e7ids of the somite, 

 the sensory (middle) ring being the stable component of the somite in 

 changes of either sort. 



The final result of somite reduction in Glossiphonia (Fig. &, A) points 

 to a primitive condition of the leech somite, in which it consisted, like the 

 somite of a choetopod, of a single ring. A tendency to increase the 

 number of annuli in the somite would, in harmony with what we know 

 of the process of ring multiplication in the leeches, have called for the 

 formation of a new and narrower ring at one end or the other of the so- 

 mite, or at both. If new rings were formed simultaneously at both ends 

 of the primitive one-ringed somite, we should arrive, by a single step, at 

 substantially the condition of somite found in Glossiphonia (Fig. 6, G). 



