POLYZOA. 297 



sequently divide and form two masses. At a still later period each mass 

 divides into an anterior and a posterior part ; the former giving rise to the 

 ovary, the latter to the testis. The similarity of this mode of development 

 of the generative organs to that observed by Butschli in Sagitta, which 

 is described in the sequel, is very striking. 



ECTOPROCTA. 



Although the embryology of the Ectoprocta has been in- 

 vestigated by a very considerable number of the distinguished 

 naturalists of the century, many points connected with it still 

 stand in great need of further elucidation. The original nature 

 of the embryo was rightly interpreted by Grant, Dalyell and 

 other naturalists, but it was not till Huxley demonstrated the 

 presence of both the ovary and testis that the true sexual origin 

 of the embryo in the ovicells became an established fact in 

 science. The recent memoir of Barrois (No. 298), though it 

 contains the record of a vast amount of research, and marks a 

 great advance in our knowledge, still leaves a great number of 

 points, both with reference to the early development and to the 

 larval metamorphosis in a very unsatisfactory condition. 



Four larval forms can be distinguished, viz. 



(1) A larval form which with slight modifications is common 

 to all the genera of the Chilostomata (except Membranipora and 

 Flustrella) and of the Ctenostomata. 



(2) A bivalved larva of Membranipora known as Cyphon- 

 antcs, the true nature of which was first recognized by Schneider 

 (No. 322), and the closely allied larva of Flustrella. 



(3) The typical Cyclostomatous larva, for the first full 

 description of which we are indebted to Barrois (No. 298). 



(4) The larva of the Phylactolaemata. 



Chilostomata and Ctenostomata. As an example of the 

 first type of larvze, Alcyonidium mytili, one of the Ctenostomata, 

 may be conveniently selected for description, as having been 

 more completely worked out by Barrois than perhaps any other 

 form. The segmentation commences in the normal manner by 

 the appearance of two vertical furrows followed by an equatorial 

 furrow, which divide the ovum into eight equal spheres. The 

 stage with eight spheres is followed, according to Barrois, by one 

 with sixteen, formed in a remarkable manner by the simultaneous 



