CH^ETOPODA. 333 



developed after the establishment of the segments, and thus 

 seem rather to be secondarily adapted to the segments than 

 primarily indicative of them. 



In most Polytrochae the rings are incomplete, so that lliey 

 fall under the category of Nototrochae or Gasterotrochae. 



The larva of Odontosyllis is an example of the former, and that of 

 Magelona of the latter. The larvae of Nerine and Spio, already quoted as 

 examples of an unsegmented arrangement of the ventral ciliated half rings, 

 are both amphitrochal forms. 



As an example of a polytrochal form with complete ciliated rings Oph- 

 ryotrocha puerilis may be cited. This form, discovered by Claparede and 

 Metschnikoff, develops a complete ciliated ring on each segment : and the 

 pras-oral ring, though at first single, becomes at a later period divided into 

 two. This form is further exceptional in that the ciliated rings are persistent 

 in the adult. 



The unimportance of the character of the rings in the polytrochal forms 

 is shewn by such facts as the absence of these rings in Terebella nebulosa 

 and the presence of dorsal half rings in Terebella conchilega. 



The mesotrochal forms are the rarest of Chaetopod larvae, 

 and would seem to be confined to the Chaetopteridae. 



Their most striking character is the presence of one or two complete 

 ciliated rings which girth the body between the mouth and anus. The 

 whole body is further covered with short cilia. The anus has a distinct 

 dorsal situation, while on its ventral side there projects backwards a peculiar 

 papilla. 



The total absence of the typical prae-oral and of the peri-anal 

 bands separates the mesotrochal larvae very sharply from all the 

 previous types. 



A characteristic of many Chaetopod larvae is the presence of 

 a bunch of cilia or a single flagellum at the apex of the prae-oral 

 lobe. The presence of such a structure is characteristic of the 

 larval forms of many other groups, Turbellarians, Nemertines, 

 Molluscs, etc. 



In the preceding section the mode of multiplication of the 

 segments has already been sufficiently described 1 . 



1 It has been insisted by Semper (No. 355) that certain of the anterior segments, 

 belonging to what he regards as the head region in opposition to the trunk, become 

 interpolated between the trunk and the head. The general evidence, founded on ob- 

 servations of budding, which he brings forward, cannot be discussed here. But the 

 special instance which he cites (founded on Milne-Edwards's (No. 347) observations) 



