32 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 



I was first enabled to determine as sensory the rings indicated by Arabic 

 numerals in the right half of Figure 4 ; further study revealed the presence of 

 marginal sensillse in the positions indicated in Figure 3. 



The metamerically repeated sensory annuli were thus positively identified 

 throughout the greater part of the body. It remained merely to mark off the 

 somite limits between successive sensory annuli. This I at first did after the 

 usage of Whitman ('85, '92) and practically all others since the time of Gra- 

 tiolet ('62), considering the sensory ring as occurring at the anterior end of 

 its somite. 



I found, however, that a consistent following of this practice would, toward 

 either end of the body, place the somite limits in the middle of a ring instead 

 of between rings, the position in which somite boundaries fall in other regions 

 of the body. See Figure A, xxv'., xxvi'., etc. 



XXlll.' i. 



XXV.' <' ^^-r ^-^_V\ 



xxvii.' -'.::: \5l.^^_^^^L >v_<4^L_A xxvii. 



Figure A. — G. stagnalis. Dorsal view of posterior part of body, showing mar- 

 ginal sensillae. Somite limits are indicated correctly at the right of the figure 

 (xxiv. to XXVII.) ; at the left of the figure (xxiii'. to xxvii'.) tliey are shown 

 as they have been commonly but incorrectly placed. 



This led me to inquire whether the sensory ring really is the anterior ring 

 of its somite. The results of this in([uiry have been publislied elsewhere 

 (Castle, 1900), so that only one or two of the more important conclusions 

 need be restated here. One of these, already suggested in part on page 29, 

 is the following : — 



Somite limits coincide with neuromeric limits ; consequently in Glossiphonia the 

 sensory ring is the middle, not the anterior ring of the somite. 



