ON THE BRITISH FOSSIL CORALS. 127 



The septal development of Heliolites is exaggerated in Propora, a genus 

 from the Upper Silurian, and which perhaps lasted into the Carboniferous. 

 The costaj in this genus are well developed, but the ccenenchymal cells are 

 less geometric than in Heliolites. The structural relations are of the closest, 

 and the generic distinction is not of the usual generic value. Another 

 Upper Silurian genus, Lyellia, represents these symmetrical Milleporidse 

 in America. The corallite walls are subcostulate and not so costulate as in 

 Propora. The septa (12) are well developed, as in Heliolites and Propora and 

 Heliopora, and the coenenchyma is perfectly vesicular — spongy, in fact, like 

 Heliopora. Here, then, in the distant and British and Northern European 

 Silurians, there were closely allied forms varying amongst themselves, but 

 more than the secondary types, the variation having some sort of likeness in 

 both instances. It is impossible not to acknowledge the genetic affinities of 

 all these genera except Millepora, of which more will be said, or to hesitate 

 to assert that there has never been a break in the Tabulata, and that the Ee- 1 

 cent and Palaeozoic Heliopora and Heliolites are very closely allied, the one 

 being the descendant of the other*. Axopora is a tertiary genus, and its 

 immense columella, which fills up the corallite inferiorly and leaves but little 

 room in the calice around it, of course prevents the tabiilae from reaching 

 across the axial space. The tabulae come in contact with but do not perfo- 

 rate the columella, so that this structure grows progressively without any 

 reference to them ; they do not form floors upon which a columella is deve- 

 loped f. There are no septa, and the coenenchyma is reticulate in the ex- 

 treme. No living analogue of this genus exists, and exception may be taken 

 whether it be a true coral. It has no Palasozoic representatives. 



Battershyia is a very remarkable Palaeozoic genus, and has been examined 

 by me+. MM. Milne-Edwards and Jules Haime§ classify it with the Mil- 

 leporidaj, but apparently only provisionally ; but it will be noticed elsewhere. 

 I have associated Battershyia and Heterophyllia together as a new division 

 of the Aporosa of the Astraeidae, under the name of the Palastraeaceae, which 

 are noticed in the first part of this Eeport. 



The Eavositidae are divided by MM. Milne-Edwards and Jules Haime into 

 the following subfamilies : Eavositinae, Chaetetinae, Halysitinae, PociUoporinae. 

 All are presumed to present the following family characteristics : — " The 

 corallum is formed essentially of the lamellar walls of the coraUites, and 

 possesses hardly any or no coenenchyma. The visceral chambers are divided 

 by tabulfe, which are numerous and well developed." 



The subfamilies without any coenenchyma, and those whose coralHt€s fonn a 

 massive corallum, are the Eavositinae and the Chastetiase, and the genera whose 

 corallites are not united on all sides the Halysitinae. The Pocilloporinfe 

 constitute the ccenenchymal subfamily. One of the great difliculties of the 

 zoophytologist appears strongly enough whilst investigating these Tabulata, for 

 the question constantly arises, and can only be answered very unsatisfactorily, 

 are such and such forms really Actinozoa? are they not Polyzoa, Hydrozoa, 

 or of some class which has become extinct, and which has no modern repre- 

 sentatives ? 



Some genera are characterized by the absence of septa. Thus Chcetetes 

 has long basaltiform coralKtes, numerous tabulae which do not correspond in 

 their- plane throughout the coraUum,. no septa, and the reproduction is fissi- 

 parous. I 



* See Huxley's Address, Geol. Soc. 1870. 



t Pal. Soc. Tertiary Corals, 3rd*Series, P. M. Buncan, pi. vii. figs. 11-15, 



t Phil, Trans, 1867. § Op. cit, p, 244. 



