4<)ti PROFESSOR ALLMAN ON THE RELATIONS OF THE CCELENTERATA. 



the atrial region of an ordinary hyclroid medusa, while it is distinctly expressed 

 in the gonophore of clamtella, where the free or manubrial region of the 

 umbrella is rudimental, and the whole gonophore, apart from the marginal 

 tentacles, becomes comparable to the atrial region of an ordinary hydroid 

 medusa. 



The accessory canals of Beroe run, it is true, close upon the walls of the 

 axial cavity until they leave these to throw themselves into the circular canal ; 

 but this fact cannot, in opposition to the greatly preponderating hydrozoal 

 features of Beroe be used as an argument for the actinozoal nature of the 

 Ctenophora. 



The accessory canals are not represented in the hydroid, while the Beroe 

 further differs from the hydroid in the presence of the two short aboral canals, 

 by which the aboral end of its somatic cavity communicates with the outer 

 world (fig. 12, t), as well as in the disposition of its so-called nervous system 

 and sense organs, and in its characteristic bands of vibratile lamellae (fig. 13, y); 

 all which features are among the special characteristics of the order, and in no 

 way justify the absorption of the Ctenophora into the Actinozoa. 



In this attempt to determine the true affinities of the Ctenophora, I have 

 taken Beroe, instead of Cydippe or other ctenophorous genus, as the subject of 

 comparison, not only because Beroe is a typical ctenophorous form, and of com- 

 paratively simple structure, but because I have myself made its anatomy and 

 development a subject of special study."' 



* Proc. Roy. Soc. Edinb., 1862. 



