CCELENTERATA. 179 



This ancestor was perhaps common to the Turbellarians also. The constant 

 presence of thread-cells in the inner layer of their epiblast fits in with their 

 derivation from a form similar to the planula. While the solid parenchy- 

 matous digestive canal of Convoluta and Schizoprora and other forms 

 amongst the Turbellarians, though very probably secondary, may perhaps 

 be explained by such a view of their origin. 



The planula in its primitive condition is not bilaterally symmetrical, but 

 frequently, as amongst the Actinozoa, it becomes flattened on two sides before 

 undergoing its conversion into the adult form. Perhaps the bilateral form 

 of planula is the starting point both for the Ccelenterata and the Turbellaria. 

 In this connection the peculiar unilateral development of a tentacle in 

 Scyphistoma and Actinia should be noted. 



The planula occurs in the majority of sessile forms of Hydro- 

 zoa except the Tubularidae and Hydra. It is also characteristic 

 of the Trachymedusae and Siphonophora. Amongst the Acras- 

 peda it is also present, but has an exceptional mode of ontogeny 

 which is discussed in connection with the germinal layers. 



It is characteristic both of the Octocoralla and Hexaco- 

 ralla, but is not found in the Ctenophora. 



In the Tubularidae and in Hydra an abbreviated develop- 

 ment leads no doubt to the absence of a free planula stage, and 

 the absence of a larval form amongst the Ctenophora may, as has 

 already been stated, be probably explained in the same way. 



The Ccelenterata of all the Metazoa are characterized by the 

 greatest simplicity in the arrangement of their germinal layers ; 

 and for this reason very considerable interest attaches to the 

 mode of formation of the layers amongst them. Two germinal 

 layers are constantly found, which correspond in a general ivaj> 

 to the epiblast and hypoblast. It might have been anticipated 

 that a certain amount of uniformity would have existed in the 

 mode of formation of the layers. This however is not the case. 

 In perhaps the majority of forms they become differentiated by 

 a process of delamination, but in a not inconsiderable minority 

 the two layers owe their origin to an invagination. 



Delamination is constant (with the doubtful exception of 

 some Tubularidae) amongst the Hydromedusae and Siphono- 

 phora. It is perhaps in the main characteristic of the Actino- 

 zoa. 



Invagination by embole takes place, so far as is known, con- 

 stantly amongst the Acraspeda and frequently amongst the 



12 2 



