222 TORREY 



factors which underHe differentiation in development, I think we 

 should suspend our judgment in regard to the significance of 

 these cells. It should also be borne in mind that these small 

 cells have never, so far as I know, been proved to be the funda- 

 ment of a typical mesoblast organ, but have been considered 

 mesoblast merely because of their position. 



The coelomesoblast cells, which at first lie closely pressed to- 

 gether, under the posterior lip of the blastopore (PI. II, Fig. 24), 

 exactly as in Etipomatiis and Podarke, soon separate and move 

 apart toward the sides (Text-Fig. 9, A^ B, C). The pressure 

 of the entoblasts causes them to lie well up toward the proto- 

 troch. They may be distinguished without difficulty from the 

 fact that their protoplasm is much more finely reticulated than 

 that of the entomeres, and is also comparatively free from yolk- 

 granules, as has been noted by Conkhn in Crepidula. The 

 lateral migration of the right cell is a little more rapid than that 

 of left, owing to the fact that it is less obstructed by the ento- 

 blast mass. When these two cells are about half way to the 

 future ventral side, each divides equally, with the plane of the 

 division parallel to the prototroch. At length, after three more 

 divisions in the same plane, the two coelomesoblast bands of 

 five cells each come to lie in the position already described. 

 These rows of cells (Text- Fig. 9, D) are thus shown to have 

 the same origin and early history as the mesoblast-bands in 

 other annelids. To this we may add the statement of Conn 

 that, during the post-larval development, these bands in TJialas- 

 sema increase in size, become segmented, and finally give rise 

 to the secondary body cavity in a typical annelid way. The 

 meager development of the coelomesoblast in the trochophore 

 is clearly correlated with the long duration of its free-swimming 

 and (as far as development is concerned) almost stationary larval 

 existence. 



It is a fact, as far as I know without exception, that in all 

 forms where there is a trochophore stage of long duration (as 

 is the case in all annelids with equal cleavage), the two coelo- 

 mesoblast cells do not, in the earlier stages at least, bud like 

 teloblasts. This is certainly true in Hydroides and the Newport 



