Professor J. W. Gregory — On the Coral Octotr emends. 11 



present doubtful. The characteristic which offers most evidence 

 for their separation is the development of the horizontal lamellae. 

 The type species of Astrceopora, A. myriophthalma (Lam.), as figured 

 by Bernard (1896, pi. xxvi), appears to be built of tubular corallites 

 with sharply denned walls, to have no septa visible on the surface, 

 and in vertical section to have no regular synapticular platforms. 

 In A. listeri, Bern, (ibid., pi. xxviii), the synapticula are also ven*- 

 irregular, and it has nothing like the regular continuous lamellae 

 shown in Keuss' figure (1866, pi. ii, fig. 3b). Specimens of 

 Astrceopora in the geological collection of Glasgow University 

 have no synapticular platforms as regular as those in Octotremacis. 

 Of the Indian Miocene species, Astrceopora hemispherica, Dune, there 

 is unfortunately no information as to the structure of the synapticular 

 platforms ; but that species differs from Octotremacis hochstetteri by 

 the greater number of septa, for according to Duncan it has six large 

 and about six small septa. According to Bernard, however, in some 

 recent species the horizontal platforms may be so regular "that the 

 whole ccenenchyma appears to be composed of tiers of synapticular 

 floors supported by columns passing through, them ". He says this 

 structure is very variable and is especially well developed in 

 Astrceopora expansa, Briigg., and A. horizontalis, Bern. In the 

 former he describes the ccenenchyma as consisting of nearly con- 

 tinuous horizontal floors (1896, p. 86). 



Thanks to the kindness of F. J. Bell, Esq., I have had the 

 opportunity of examining the types of these two species. A. expansa 

 has a flat corallum, shaped like a wide incomplete funnel ; at one 

 part of the base a vertical section £-, in. in thickness shows that 

 the horizontal lamellae are continuous and regular. The type- 

 specimen of A. horizontalis, Bern., is a thin sheet, convex above and 

 concave below. Unfortunately the edges are nearly perfect, so that 

 they do not show the internal structure, and the fractures on the 

 under surface do not give any clear evidence as to the nature of the 

 horizontal lamellae. As the type is the only specimen, it is thought 

 unjustifiable to cut it in order to determine its structure. 



If Bernard's view be correct, the horizontal lamellae are of no 

 diagnostic value, as they depend largely on the shape of the corallum, 

 in flat coralla the vertical element being reduced and the horizontal 

 platforms being well developed. The spreading growth of some 

 coralla may, however, be the result rather than the cause of the 

 predominance of the horizontal element. It does not seem possible 

 to decide as to the value of this character until the internal structure 

 of the recent species is better known. The most convenient course 

 at present is therefore to retain Octotremacis provisionally as a sub- 

 genus of Astrceopora characterized mainly by the regular development 

 of its synapticular platform. 



Age. — Octotremacis was identified as Miocene because the other 

 •corals from Java described by Reuss were of that age. The type of 

 this genus was, however, found under different conditions from the 

 rest of the collection. It was a fragment collected from a trachyte 

 breccia at Tjukang-Raon, and there is no direct evidence from Java 

 as to its age. It may have been introduced into Miocene tuffs 



