xv AFFINITIES OF THE GROUP 447 



rather primitive features, and so is the Platyhelminthine character 

 of the excretory organs. With regard to the vascular system, 

 however, it should be pointed out that it arises very late in the 

 larva of those Gephyrea whose development is known, and that 

 it does not seem to correspond with the vascular system of other 

 animals; it has no fine vessels or capillaries connected with it, 

 and apparently does not act so much as the channel of the 

 circulatory medium, but more as a mechanism for the expansion 

 of the head appendages, the tentacles in the Sipunculoidea and 

 the proboscis in the Echiuroidea ; moreover, it is absent in some 

 genera of the former, such as Oncknesoma, Tylosoma, and Peta- 

 lostoma, where there are no tentacles. 



The conclusion of the whole matter seems to be that the 

 Priapuloidea are an isolated Order retaining many primitive 

 features, and having no closer affinities to the Sipunculoidea than 

 to the Echiuroidea. 



Hatschek came to the conclusion, from his work on the 

 development of Echiurus, that the Echiuroidea are true "An- 

 nelids," and from the presence and mode of formation of the 

 bristles, that they are related to the Chaetopods. In this view 

 he is confirmed by Conn, who worked out the development of 

 Thalassema. This relationship is further confirmed by the discovery 

 of Sluiter's that Sternaspis, the genus of Chaetopods which in 

 other respects most nearly resembles the Gephyrea, has in one 

 of its species (S. spinosa) a well-marked bifid proboscis, which, 

 like that of the Echiuroidea, is thrown off at the least disturbance. 

 Thus it seems fairly well established that the Echiuroidea 'are 

 closely connected with the Chaetopoda, for although the only traces 

 of segmentation they retain in the adult are the serially-repeated 

 nephridia of Thalassema and Echiurus pallasii, and the two rows 

 of peri-anal bristles in the latter, and possibly the circular nerves 

 given off from the ventral cord, yet the larva is fully segmented, 

 and in other respects is almost typically Chaetopodan. 



The relationship of the Sipunculoidea to the Echiuroidea is 

 a more doubtful point. Hatschek is inclined to separate them, 

 and in this he is again supported by Conn. Embryology unfor- 

 tunately does not help us much. The early stages and larvae 

 of Sipunculus nudus and of Phascolosoma dongatum have been 

 investigated by Hatschek and by Selenka respectively. In 

 neither genus is there any trace of segmentation or of Annelid 



