502 Mr. H. M. Bernard on the 



in open communication with one another and with the polyp- 

 cavity through pores. 



This description disposes of Milne-Edwards's distinction 

 between the genera Turhinaria and Madrepora^ that in the 

 former the coenenchyraa is distinct from the mural tissue of the 

 polyp. The description above given would, if we allow for 

 the different density of the structures, apply equally well to 

 a section through an axial polyp of a typical Madrepore. 

 " Concentric circles of thin calcareous structure are seen 

 separated by radiating linear pillars, the circles having been 

 in turn outside walls and the radii either spinules or costee " *. 



An interesting question arises as to whether these radiatino- 

 plates are true morphological costse, i. e. outward prolon^-a- 

 tions of the septa. Mr. Brook f found no connexion between 

 the so-called costse in Madrepora and the septa; and the 

 same is true of the few sections of stalks which I have been 

 able to examine in Turhinaria. In spite of this fact 

 however, I am persuaded that primitively such a connexion 

 existed, and that it has been secondarily obliterated. My 

 chief reason for believing that it was the primitive arrano-e- 

 ment is to be found in the fact that in many Turbinarians 

 the septa are directly continued into the ridges of the coeneii- 

 chyma, and that the direct connexion between the septa and 

 these costai can very often be traced in the young calicles 

 forming along the margin of a corallum (PI. XIX. fig, 2). 

 On the other hand, secondary obliteration of the connexion 

 when it ceased to have any special use might easily take place. 



The process can, indeed, perhaps be traced as follows : 



Between each pair of septa the rudiments of new cycles of 

 septa are in many cases visible. In those cases in which the 

 septa are continuous with the ridges of the coenenchyma other 

 ridges frequently run up to the edge of the polyp-cavity 

 terminating between the two septa, and are unmistakably 

 suggestive of a fresh cycle of septa. We thus have a o-reat 

 many more radiating costa3 than septa abutting on the imme- 

 diate wall of the cavity, the crowded costte representino- not 

 only the actually developed septa, but rudimentary cycles of 

 septa. If, together with this crowding of costse round the 

 polyp, we take into consideration the more or less accidental 

 variations in their thickness in the gradual process of streno-th- 

 ening the corallum, there is no difficulty in understandino- 

 how the primitive connexion between the costse and the septa 



* Martin Duncan, " On the Hard Structure of some Species of Madre- 

 pora,^^ Ann. & Mag. Nat. Hist. 1884, vol. xiv. p. 191. 



t " The Genus Madrepora,^^ vol. i. of the British Museum Catalou'ue of 

 the Madreporaria, 1893, Introd. p. 9, 



