10 Professor J. W. Gregory — On the Coral Octotr •emacis. 



structures. If the coral had undergone a double change in fossiliza- 

 tion Reuss' view was possible ; and his statements that the circular 

 rods are casts of canals were so positive that, though in 1900 

 I mentioned my doubts, I felt bound to dismiss them. So I accepted 

 the fossil as a Miocene Helioporid, and as the name Polysolenia was 

 preoccupied renamed the genus Octotremacis (1900, p. 302). 



A fossil collected by the Right Hon. Sir William Macgregor, 

 G.O.M.G., C.B., etc., from the upper part of the Ely River in New 

 Guinea, shows that my suspicions were justified. This specimen 

 seems to me almost certainly the same species as Polysolenia 

 hochstetteri. It agrees very closely with the figures by Reuss, the 

 only noteworthy difference being that the horizontal synapticula are 

 less regularly horizontal than they are there represented. First 

 inspection of the transverse section suggested that in the New 

 Guinea specimen the corallites are considerably larger than in those 

 from Java ; but this appearance is due to the section having been cut 

 obliquely. The tubes cut at right angles and the shorter width of 

 those cut obliquely have dimensions very slightly larger than those 

 in Octotremacis hochstetteri. In the type of the species the corallites 

 have a diameter of 2 mm., and are from 1\ to 4 mm. apart ; whereas 

 in Sir William Macgregor's specimen the diameter is from 2 to 

 2\ mm. and the corallites are from 2£ to 5 mm. apart. There 

 seems no difference between the corals adequate for their specific 

 separation. 



This Fly River coral has, however, clearly a trabecular 

 ccenenchyma, which is composed of parallel vertical rods. Some 

 of the rods project slightly into the tubes of the corallites, and are 

 suddenly reduced in thickness to very thin irregular septa. This 

 specimen shows that the thick vertical structures identified by Reuss 

 as the septa are casts of the interseptal spaces. The septa are the 

 thin intervening lamellae. According to Reuss their number is 

 eight; but he shows one corallite with seven, and he refers to the 

 occurrence of only six septa. In Sir William Macgregor's specimen 

 the number six is the more common, though specimens with seven 

 and with less than six septa also occur. 



As this coral has a trabecular and not a tubular ccenenchyma, its 

 systematic position has to be changed. Instead of being an 

 Heliolitid its affinities are with Astr&opora. It agrees with that 

 genus in its trabecular coenenchyma, circular calices, the extreme 

 thinness of the septa, which are absent from some corallites, and the 

 presence of horizontal laminae, which in vertical sections, as is well 

 shown in Reuss' figure, pi. ii, 3^, have almost the aspect of tabulae. 

 This lamellar structure occurs in many Fungians owing to the 

 development of synapticular platforms. Octotremacis may therefore 

 be diagnosed as follows : Madreporidae with a massive corallum and 

 a loose ccenenchyma composed of vertical trabecula and regular 

 conspicuous horizontal synapticular laminae. The septa are very 

 thin (but better developed than in Astrceopora) ; they vary in number 

 usually from six to eight. No columella. Type, Octotremacis 

 hochstetteri, Reuss sp. Eocene, Java. 



Whether Octotremacis is generically distinct from Astraopora is at 



