6 DISCOPHOR^. Part III. 



while it gives a more definite form to the digestive cavity and keeps its lower 

 floor permanently apart from the upper floor, also secures a greater independence 

 to the apparatus corresponding to the proboscis of the CryptocarpJB as a whole, 

 and to each of its four prominent appendages separately. Moreover, the side walls 

 of the digestive cavity are comparatively thin in the intervals between the four 

 pillars; so much so, indeed, that the walls appear perforated, and have generally 

 been described as perforated, when in reality these seeming holes are walled over 

 by a veil more or less tightly stretched across the holes, and frequently forming 

 pendent pouches, to the inner surface of which the ovaries and spermaries are 

 attached. The chymiferous system never consists of simple tubes immediately arising 

 from the central cavity and reaching directly a simple circular tube, but always 

 presents complicated anastomoses at the margin of the disk, while the channels 

 arising from the central cavity are either simple tubes or wide sacs opening freely 

 into the main cavity. As in the Cryptocarpte, the tentacles are either few occu- 

 pying a special position, or many along the whole margin of the disk, or they are 

 entirely wanting. The eye-specks are always at the peripheric end of simple 

 radiating tubes, but never at the base of a tentacle along the circular tube. They 

 are frequently, but not always, protected by folds of the margin of the disk. 

 The margin of the disk is very thin, and sometimes turned inward, in the shape 

 of a veil. 



The position of the ovaries and spermaries is so peculiar, and contrasts so strik- 

 ingly with that of the Cryptocarpoe, that the way in which the eggs are freed 

 is very different. In the Phanerocarpre the egg sacs are arranged in loops or 

 festoons upon the inner surface of the veil closing the lateral holes or pouches 

 of the main cavity, and when the eggs are detached they move into the digestive 

 cavity, and, following the channels formed by the ann-like prolongations of the four 

 pillars which support its angles, finally reach the little marginal sacs of the arms, 

 in which they remain until they are cast ofi" into the surrounding medium. Peculiar 

 as this structure may seem when compared to that of the Cryptocarpae, there is 

 yet the closest homology between them, for the large pouches containing the ovaries 

 and spermaries of the Phanerocarpte are, after all, only dilatations of the chymiferous 

 system along the course of its radiating channels; while in Cryptocarpte, instead 

 of large pouches there are simple, narrow tubes, upon the sides of which the eggs 

 are developed, and from which they immediately drop into the surrounding medium. 

 The fact, that in some CryptocarpsB the eggs are developed upon the proboscis, 

 in no way conflicts with this explanation, since the angles of the proboscis, as may 

 best be seen in Bougainvillea, are quite as much the direct prolongation of the 

 radiating tubes, as the ovarian pouches of Cyanea are a direct prolongation of its 

 radiating chymiferous sacs. 



