POLYZOA. 563 



parts in the adult (see p. 552). On the other hand, it may be 

 argued with equal force that the oro-anal surface of the adult is 

 equivalent to the dorsal surface of the larva, for the larva attaches 

 itself by a sucker on the ventral surface between the rtfbjith and 

 anus. In this case we should have to regard the formation of the 

 first polypide as a case of larval metamorphosis, comparable to that 

 of a holometabolous insect, in which the changes are so extensive 

 that many of the larval organs degenerate completely, and are re- 

 formed in the imago. If this view is correct the reduction of the 

 dorsal surface and the approximation of the mouth and anus in 

 the adult as compared with the larva, would be comparable to the 

 phenomena which take place in Actinotroclia at the metamorphosis 

 into the adult Phoronis. On the whole we are inclined to the view 

 that the first polypide is formed by budding from the primary 

 zooecium, which is the metamorphosed larva, and that consequently 

 the ganglion of the Polyzoa is really a ventral, suboesophageal 

 ganglion, and that the surface between the mouth and anus of the 

 adult is the ventral surface, and comparable to the surface on which 

 the mouth and anus of the larva are placed. In coming to this con- 

 clusion we attach great importance to the fact that the formation of 

 the first polypide from the primary zooecium is closely similar to the 

 budding which takes place in the formation of the adult colonies. 



Before leaving the subject of the development of the Ectoprocta, 

 attention must be called to a most important process of embryonic 

 fission discovered by Harmer in the Cydostomata. In this sub-order 

 the ovicell is a zooecium with a rudimentary polypide (Crisia), in 

 the funicular tissue of which the ovum arises. Fertilisation has 

 not been observed, but the alimentary canal of the polypide grows 

 round the egg and forms a multinucleated follicle for it. The 

 ovum segments and becomes a multinucleated mass, in which cell 

 limits cannot be seen. This buds off small pieces, which become 

 ciliated, escape from the zooecium in a manner not clearly ascer- 

 tained, and swim away as ciliated larvae. 



Asexual reproduction by budding is of course a most important 

 process in the Edoprocta. As shown above, it begins even in the 

 larva, and is a marked feature of the metamorphosis ; and the forma- 

 tion of the colonies is entirely due to it. Indeed, we may say that 

 the Edoprocta resemble the Siphonophora and other Coelenterates in 

 possessing that power of multiplication and of differentiation of parts 

 (e.g., avicularia, etc.) which is so preeminently a characteristic of the 

 vegetable kingdom. 



