ON THE BRITISH FOSSIL CORALS. 119 



I do not consider that the Tubulosa belonged to the Madreporaria, but 

 that they were Mcyonarians. 



It is very certain that some Aporose, Perforate, and Rugose corals have 

 tabulag, and that their existence cannot remove the forms from their re- 

 ceived zoological position into the separate section of Tabulata. 



Thus the well-known Aporose coral of the deep sea, Lopliohelia pro- 



move, than portions of Stromatojiora which enclose the corallites and grow simultaneously 

 with them. 



I have altered the generic characters of Battershyia, in consequence of a careful exami- 

 nation of the old and the two new species. It is as follows : — Corallum fasciculate and 

 branching ; corallites tall, cylindrical, unequal in size and distance ; septa numerous and 

 following no apparent cyclical order. 



Endotheca very abundant : it is vesicular, and there are no tabulre. Epitheca, costa?, 

 and coenenchyma wanting. The wall is stout, and the septa spring from wedge-shaped 

 processes. The columellary space is occupied by vesicular endotheca. Gemmation extra- 

 calicular and calicular from buds having only five septa. 



There are three species : — 



Battersbyia inrequalis, I>uncan. I Devonian Limestone ; 



grandis, Duncan. \ found in pebbles, 



gemmans, Duncan. J and not in situ. 



In Battershyia gemmans the buds which develop more than five septa grow into coral- 

 lites, which are destined to bud again from the external wall, and the buds which de- 

 velop five septa produce other buds from their interseptal loculi ; the buds thus developed 

 resemble the corallites with more than thi-ee septa. This curious alternation of gemma- 

 tion has not been noticed in any other genus. 



The genera Battershyia and Heterophyllia (Phil. Trans, he. cit.) have much in common. 

 They have a stout wall, a vesicular and dissepimental endotheca, delicate septa, very irre- 

 gular in their number, and neither tabular epitheca nor a quaternary septal arrangement. 



The genus Batfcrsbyia has nothing to ally it to the Eugosa or the Tabulata. Hetero- 

 fhyllia lias in some of its species the solitary septum or vacancy whicli is so often observed 

 in the Cyathophyllida^. Its costal wall and endotheca connect it with the Mesozoic and 

 recent Astrseidoe. 



The singular septal development of Battershyia is witnessed in the fasciculate Liassic 

 Astrajidse. The pentameral arrangement of the Battersbyian septa is not unique, for 

 Acanthoeoenia Bathieri, D'Orb., of the Neocomian has only five septa, and so have the 

 species of Pentaccenia, all of which are from the same great formation. The proper Liassic 

 and some of the Lower Oolitic Thecosmiliaj and CalamophylUiB represent and are allied 

 by structure to Battershyia. The highly specialized characters of the Heterophyllise, espe- 

 cially of H. mirabilis, could hardly be perpetuated during great and prolonged emigra- 

 tions, so that the genus appears to be without representatives in the secondary rocks. Its 

 alliance to Battershyia, however, is evident enough. 



The genus HeterophylUa, MCoy, was examined by me in 1867, and the study of several 

 new species of it rendered a fresh diagnosis requisite. 



The following description of the diagnosis appeared in my essay on the genera Hetero- 

 phyUia, &c., already noticed: — 



"The corallum is simple, long, and slender. The gemmation takes place around the 

 calicular margin, and is extracalicular. The septa are either irregular in number and 

 arrangement, or else are six in number and regular. The cost« are well developed, and 

 may be trabecular, spined, and flexuous. The wall is thick ; there is no epitheca, and the 

 endotheca is dissepimental." 



The genus may be subdivided into a group with numerous septa, and a group with six 

 septa. 



In the first the rugose type is faintly, and in the last the hexameral arrangement is well 

 observed. 



The dense wall and the dissepimental endotheca prove that the type of the Mesozoic 

 Coral-fauna was foreshown. 



The endotheca varies in quantity in the different species, and it resembles the tabular 

 arrangement ; but even when this is the ease and the cross structures are well developed 

 and numerous, they do not stretch over the axial space, so as to shut out cavities 

 as if they were floors ; they do not close in the whole of the visceral and interlocular 



