MADBEPORAEIAN SUBFAMILY LOPHOSERINJE. 317 



The corallum is compound and more or less foliaceous, and tlie 

 basal wall is perforated like that of the Funginse. There are no 

 synapticula, and there are endothecal dissepiments. 



An examination of specimens of two species proves that 

 this perforation of the wall is not invariable, that it does not 

 occur in some collines at all, and that in otbers, where it does 

 occur, it has not the same significance as the perforations of 

 the wall described in the first part of this communication (see 

 Journ. Linn. Soc. vol. xvii. pi. v. figs. 1, 2, 3, and pi. vi. figs. 3 & 6). 



In some old specimens the foliaceous expansion is only seen at 

 the margins of the corallum, and elsewhere there are calices and 

 collines on both sides of the colony. There is no perforation in 

 such instances ; and the small foramina seen between the septo- 

 costge open into a dissepimental cell-space far above the wall. 



Moreover, the importance of all synapticula in relation to 

 the interloculi and their invariable presence in ITungidse takes 

 the Merulina quite beyond that family. I relegate it to th.e 

 AstrseidsB, especially as the transitional nature of Siderastrcea is 

 so evident. 



The genus Echinopora of Lamarck and Dana was very naturally 

 called Agaricia by Schweigger, for it has a very Eungid appear- 

 ance. Yerrill has placed the genus amongst tbe Fungidae ; and 

 an examination of some specimens proves that he is correct j 

 but I fail invariably to find synapticula, and, indeed, where I have 

 detected them, tbey are very ill-developed and near the bottom of 

 the interseptal loculi, and amongst much dissepimental endotheca. 

 In some places the upper free edges of the septa fuse in a synap- 

 ticular fashion. 



VII. Remarks and Considerations regarding Classification. 



It is evident that in the genera Loplioseris, Mceandroseris, JPle- 

 sioseris,PacTiyseris, and Coscinarcea the synapticula are important 

 structures, not exaggerated granules, and that they may be the 

 only means by which the septa are united and the cohesion of the 

 corallum maintained above the solid basal wall. It is evident that 

 tbey may fuse and form a false wall which may limit calices and 

 unite the septa along collines. I do not distinguish in the recent 

 species of th.e subfamily Lophoserinse any structures which may 

 be called false synapticula. 



The synapticula occupy much of the interseptal loculi, and 

 tend to produce canal-shaped spaces therein, more or less vertical 



LINN. JOURN. — ZOOLOGY, YOL. XYII. 23 



