CORALS FROM THE PERSIAN GULF. 1041 



satisfied that these characters — the octoradiate symmetry, the 

 pronounced accretion ridges and the septal fusions in the lower 

 part of the coral — really indicate any true affinities with Palfeozoic 

 corals. Duncan found one specimen of Guynia in which, owing 

 to the abortion (?) of two septa, the upper part of the corallum 

 showed a hexaradiate symmetry. This specimen Avas regarded 

 by him as "very suggestive in the matter of the evolution of 

 the hexameral from the octomeral types, or rather from the 

 tetrameral." But the view that was in his mind, that the 

 modern hexaradiate coi-als are descended from a previous tetra- 

 radiate type, is not one which commends itself to more recent 

 investigators of coral structure. The researches of Duerden* 

 suggest that the tetraradiate symmetry of Rugose corals is derived 

 from a more primitive hexaradiate symmetry, and there is no 

 evidence in the development of modern corals, or satisfactory 

 evidence in palteontology, that the modern hexaradiate corals ai"e 

 derived from a previous tetraradiate or octoradiate ancestry. 



It is true that many Palseozoic corals exhibit fovir dominant 

 primary septa, but there are none with eight dominant septa. 

 The true octoradiate condition is known only in certain species of 

 the Tertiary genus Conosmilia and in the recent genera Guynia 

 and Pyrojihyllia. The recent genus Haploi:)hyllia has also eight 

 primary septa and eight secondary septa, but it is suggested by 

 Gardiner that this genus is only a growth stage of Dunccmiia. 

 The close relationship of Co7iosmilia to the simple modern 

 hexaradiate corals suggested by the classification of Duncan and 

 of Ogilvie, seems to point to the conclusion that the Guyniidfe 

 represent an early offshoot of the modern hexaradiate line of 

 descent, and that their true affinities lie rather with the modern 

 Turbinoliidse than with aiiy known Palseozoic coral. 



I may refer in conclusion to some remarks made in my 

 preliminary paper on the presence of endotheca in Pyrophyllia. 

 Although Duncan's definition of endotheca is not very explicit, 

 I have come to the conclusion that there is no structure in 

 Fyrophyllia that really corresponds with what is u^sually regarded 

 as endotheca by the palseontologists. A longitudinal section of a 

 Pyrophyllia (PL LVIII. fig. 19) shows three regions: an upper 

 calicular region {x) in which the septa are free from one another 

 and from the columella, a middle I'egion [y) in which the septa 

 are more or less irregularly fused with one another and with the 

 columella, and a lower region (s) in which the septa are again free 

 and the thecal wall infiated. 



It seems probable from the texture of the well-preserved 

 specimens that the living tissues were confined to the upper 

 calicular region. The growth in thickness and the fusion of septa 

 and columella in the middle region (see text-fig. 221) perform the 

 same function as the endotheca of Palaeozoic corals, as they " unite 

 septa, close the loculi, and enable the coral to grow in height 

 and strength, and limit the growth downwards of mesenteries 

 * Ann. N. H. xviii. 1906. 



