NOTES ON ACTINOTROCHA. 477 



body-wall some distauce posteriorly to the bases of the 

 tentacles, and runs forwards and inwards to meet tlie wall of 

 the gut. There is a well-marked epithelium lining the 

 cavity. Dorsal and ventral mesenteries do not appear to be 

 present. The muscle strands above mentioned ran through 

 this cavity. Each divides into two bands, one of which runs 

 through the collar cavity and meets the dorsal body-wall, and 

 runs posteriorly along one side of the median line ; the other 

 runs ventrally, externally to the oesophagus, meets the 

 ventral body- wall and then runs posteriorly along one side 

 of the median line. The collar cavity, which is well developed 

 in the young larva, becomes, in the later stages, obliterated 

 by the development of a large mass of blood- corpuscles. 

 These are developed around the ventral diverticulum of the 

 stomach, and in the old Actinotrochse often fill the collar 

 cavity. There is no epithelium between these corpuscles, 

 which form a mass adhering to the wall of the stomach and 

 the collar cavity, and there cannot be any doubt that they lie 

 freely exposed to the collar cavity (tig. 6). This explains 

 Caldwell's statement, often disputed, that free communication 

 exists between the body-cavity in front of the septum (i. e. 

 the collar cavity) and the vascular system. The blood- 

 corpuscles do not, however, float in the fluid in the collar 

 cavity, so that there is no communication between the collar 

 cavity and the vascular system in the sense that blood flows 

 into the collar cavit}' and mixes with the ccDolomic fluid. The 

 mass of corpuscles is connected with the dorsal blood-vessel 

 Avhicli runs along the dorsal wall of the stomach posteriorly. 

 The mass of corpuscles breaks up during metamorphosis, and 

 the corpuscles are apparently withdrawn into the blood- 

 vessels. With regard to the origin of the corpuscles, it may 

 be said that they are formed from the splanchuopleure (fig. 5), 

 covering the stomach and its diverticulum, the cells of which 

 proliferate and become blood-corpuscles. Blood-corpuscles 

 are also found between the splanchnopleure and the stomach 

 Avail, having been found, apparently, in the same manner. 

 It may be mentioned here that, according to Cori, blood- 



