222 . CEYLON PEARL OYSTER REPORT. 



palial and columellar structures. As may be seen in fig. III., C, the hammer-shaped 

 palial jii'ocess at the inner end of the primary septum cuts off the filaments of the 

 primary mesenteries from the axial space, and in D the palus is seen to have contracted 

 unions with the columellar pillars, cutting off the interseptal loculi from the axial 

 space. It is evident that the broad ciliated filaments, extending far down in these 

 nearly isolated loculi, are the chief if not the only agents in maintaining the circulation 

 in the deeper parts of the loculi. Fig. III., F, shows how much the interseptal loculi 

 become narrowed and isolated in the deeper parts of the coral in consequence of the 

 abundant secondary thickening or stereoplasm. 



All the mesenteries are fertile. In female jDolyps the ripe ova are large, and filled 

 with granules of deutoplasm. I have been unable to make out details to my 

 satisfaction, but the ova, when young, appear to become enclosed in the mesoglcea, 

 and as they increase in size they project from the sides of the mesenteries, still 

 enclosed in a thin mesoglceal envelope, outside of which is a layer of endoderm, forming 

 a sort of follicle. The ripe ova are pyriform and hang in bunches from the sides of 

 the mesenteries, each ovum attached by a slender stalk of mesogloea and surrounded 

 by its follicle of endoderm cells. 



The relation of the mesenteries to the septa affords a strong presumption in favour 

 of the view that these structures have been formed according to Pourtales and 

 Duerden's law. As has been shown, the septa are alternately exoccelic and 

 endocoelic. The exoccelic septa, which on the usual system of notation would be 

 called quaternaries, are larger than the endocoelic tertiaries, aud, as is shown in 

 fig. III., B and C, their inner ends meet and unite in front of the latter. Further than 

 this, in each system the exoccelic septa adjacent to the primaries converge and meet 

 together in front of the secondaries, forming the more or less distinct chevron-shaped 

 pali described above. This union can best be seen in fig. II. As the tertiary 

 endocoelic septa are enclosed within the smallest and therefore the most recently 

 developed mesenterial pairs, and as they are themselves the smallest and least exsert 

 of all the septa, and only unite with the Y-shaped figures formed by the exoccelic 

 septa low down in the corallum (fig. III., B, C, and D), the evidence that they are the 

 most recent in point of formation, aud that they and the mesenterial pairs embracing 

 them originated between the diverging Y-shaped outer ends of the exoccelic septa, in 

 the manner described by Pourtales and Duerden, is sufficiently convincing. 



The little canals or tubes, running inwards from the lateral walls of the coral and 

 opening into the Aspidosiphon chamber, have been described and figured by several 

 authors, but their minute characters have not yet been investigated. They are 

 almost exactly like the similar tubes in Hetaropsammia, and the transverse section, 

 Plate IV., fig. 24, of a tube of the latter genus serves equally well for Hetcrocyathus. 

 The resemblance is the more striking because in Heteroeyatlms, an imperforate 

 Turbinolid, there are endodermal canals, usually twelve in number, closely attached 

 to the whole length of the tube, which is itself lined by an invagination of the 



