568 I. IK EDA : 



blood c'orpuscle.s of Actlnoirocha do not arise at the expense of the 

 splanclinic walls, l)ut are prodnced l)y a eontinnal division of 

 certain previonsly differentiated mesoblast cells. 



I will next descrilje the blood vessels which can be seen 

 durino; larval life, with reference to the formation of the adult 

 collar cavity. I have shown in the foregoing pages that Mastek- 

 man's sulmeural shuis is prol)ably nothing ])ut a posterior recess 

 of the preoral body-cavity, and that neither the dorsal vessel 

 nor the dorsal mesentery is present on the œsophagus in any 

 species of Actinotrocha I have been aljle to obtain. I have en- 

 deavoured to ascertain the presence of the dorsal and the ventral 

 vessels as well as of the ring-sinuses around the gut, and I am 

 convinced that at no time during larval life any vessels other than 

 the dorsal on the stomach and the cu'cal capillaries are present in 

 the larvœ. 



In the .1 type-larva of 14 arms, the dorsal vessel, as figs. ÖO h 

 and 50 G will show, is not yet formed and the stomach wall is 

 uniformly lined with a thin mesoblastic layer. This layer thickens 

 later and its constituent cells become musculai', l»eginning first at 

 the base of the postoral sc})tum and along the mid-dorsal line. 

 AVhen the larva grows to the stage of 1(3 tentacles, the dorsal vessel 

 is inceptionally formed. It arises as a solid cord of cells inter- 

 posed between the nmscular, and the entoblastic, walls of the 

 stomach. As shown in fig. -32, the vessel in section is represent- 

 ed by a loose mass of mesoblast cells distinctly delimited on all 

 sides from the surrounding parts ; l)ut as yet no lumen is visible 

 in it. I have not seen the definite lumen establish itself in this 

 rudiment of the dorsal vessel at any time during the whole larval 

 life ()f this type, while, on the other hand, in the advanced larvae 

 of the other three types, it could Ijc readily recognized as such. 



