300 ME. J. E. DUEEDEN ON 



perhaps be explained as a result of the late formation of the 

 coelenteric cavity, of the assumption of the bilateral stage of 

 the tentacles before any decided internal bilateral symmetry has 

 been established in the upper tentacular region, the nearly solid 

 interior having no influence on the method of grouping of the 

 tentacular outgrowths. 



"Where, as in Arachnactis, the outgrowth of the tentacles 

 closely follows the production of mesenterial spaces, it is obvious 

 that the two must be directly related, and that the external sym- 

 metry will be moulded upon the internal. Also in the species 

 studied by Lacaze-Duthiers and others, where the ccelenteric 

 chambers were formed in advance of the tentacles, the internal 

 symmetry first assumed would be likely to impress itself upon 

 the external organs arising later. 



The irregularity in Lelrunia is probably rectified in the sub- 

 sequent rearrangement of the tentacles in hexamerous cycles 

 for, in the numerous adult specimens I have examined, the 

 tentacles, oesophageal axis, and mesenteries exhibit the relation- 

 ships usual in the Hexactiniae. 



V. COINCLTJSIONS. 



From the foregoing account it is obvious that the larva of 

 LebruQiia coralligens presents us with very unexpected conditions 

 in Scyphozoan development ; and this not alone in any one 

 particular organ, but in almost every essential structure. The 

 early tetrameral symmetry, followed by a bilateral phase, and 

 that again by the hexamerous adult ; the system of ciliated 

 ccelomic spaces connected with a closed archenteron, all embedded 

 in a mass of undifferentiated tissue ; the formation of the oeso- 

 phagus by the breaking down of the floor of an ectodermal in- 

 vagination in association with an archenteric tube ; and the 

 origin of the adult gastro-coelomic cavity from a primary coelome 

 and disintegration of the tissues, are all unique characteristics. 



The species seems to retain to a late period certain ancestral 

 characters which in other forms are either passed over or dis- 

 appear very early, so that other features dependent upon their 

 presence are never exhibited. Thus, were the vacuolated tissue 

 insufficiently developed, or to disappear early, we should have no 

 clear evidence of a distinct larval coelome. 



Some of the facts observed in the early larvae appear to obtain 



