236 CEYLON PEARL OYSTER REPORT. 



arrangement of the Eupsammiidee, which has been commented upon by many authors, 

 is due to the sequence in question. I have further shown that the septal arrangement 

 in Heteroeyathus, an imperforate coral, can only be interpreted upon the same 

 principle, and Duerden's account of the development of Siderastrcea is of itself 

 sufficient evidence that a similar mode of septal sequence prevails both in perforate 

 and imperforate corals. 



On the other hand, it is clear from von Koch's (32) and de Lacaze Duthier's (33) 

 account of the development of Caryophyllia, that Pourtales and Duerden's rule is by 

 no means applicable to all corals, but that, contrariwise, there is another and distinct 

 mode of septal sequence which is either independent of the formation of exosepta, or 

 in which the part played by the exosepta is very different. 



It is tempting to suppose that there are two principal modes of septal development 

 among corals, typified by the Eupsammiidse, Heteroeyathus and Siderastrcea on the one 

 hand, and Caryoj)hyUia on the other hand, and that the presence or absence of 

 exosepta will indicate the group to which any given coral belongs. But while I think 

 that it is very possible that a solution of many classificatory difficulties may be found 

 by extensive investigations on the lines laid down by Duerden, I must recognise 

 that it would be premature to make any positive statements in the present state of 

 our knowledge. There can, I think, be very little doubt that forms like Mussa, 

 Euphyllia, and Galaxea, which have endosepta only, cannot have been developed on 

 the same plan as Siderastrcea, but must have followed the mode described by 

 von Koch for Caryophyllia. But this statement cannot be extended to all forms 

 known to possess endosepta only. Gardiner (22) has shown that in Flabellum the 

 new cycles of septa are at first exoccelic, and that in the course of growth a pair of 

 mesenteries is formed in connection with each exoseptum, in consequence of which 

 the exosepta become endosepta. I have shown (6) that the same rule holds good 

 for the anthoblasts of Fungia. Though our evidence is not complete upon this point, 

 I believe that this will be found to be the usual mode of septal formation in all corals 

 which have only endosepta in the adult condition. It must be borne in mind that a 

 coral in the course of growth increases not only in height but in diameter. As a 

 consequence of its constantly expanding perimeter, the peripheral ends of the original 

 radial structures (the primary and secondary septa) become further and further 

 removed from one another, and the polyp forms new radial folds between which new 

 radial calcareous structures are formed to fill up the gaps between the diverging 

 septal extremities. These new radial structures are the exosepta or their equivalents. 

 In those cases in which the exoseptal formations acquire a very intimate union at the 

 time of their appearance with the adjacent endosepta, further peripheral growth is 

 accompanied by a bifurcation or splitting of the peripheral ends of the exosepta, and 

 a newer set of radial folds with corresponding calcareous structures is formed within 

 their bifurcated extremities, as in the Eupsammiidse, Siderastrcea, &c. In other cases 

 further peripheral growth is provided for by the formation of new radial folds on 



