8 GENERAL REMARKS. 



lime, excepting in the older parts. The corallite-wall is very perforate and eostate externally; 

 the costse often project as trabeculae. The wall of young corallites consists of only one layer 

 of mural tissue, and is provided externally with plain or finely serrate costae for some distance 

 down. In older corallites spinules are always present on the free edges of the costae, and 

 these are largest round the bases of the corallites, where they form a basis of lax tissue which 

 is the first stage in exogenous growth, and occasionally assists in the production of buds. 

 The formation of mural tissue takes place in the following manner : — The spinules act as 

 props on which a porous layer of new tissue is formed in such a manner that the old wall 

 forms the floor of a chamber, and the spinose costae the supports of the new roof. In old and 

 in axial corallites the costae assist in the formation of a new layer of tissue by sending out 

 lateral spinules from the free edges. In transverse sections of corallites of moderate thickness 

 the wall is seen to consist of " concentric circles of thin calcareous structure . . . separated by 

 radiating pillars ; the circles having been, in turn, outside walls and the radii either spinules 

 or costae." If the corallite is old the first, second, or third circles of tissue, next to the 

 septal cavity, are dense. The costae are imperforate, and as a rule the septa also. Buds from 

 a radial corallite are formed from trabeculae which arch over and form a low hood. The 

 costae next appear on the outer surface of the hood, and after elongation the septa appear ; 

 these arise first as linear series of spines directed inwards. No communication between the 

 cavities of the bud and axial corallite exists, " except in a very indirect manner and through 

 the medium of the dermal structures. Budding takes place remote from the calicular 

 margin, and may arise from sclerenchyma remote from the wall of the corallite." 



In the second species described, probably a variety of M. cytherea, Dana, from Mada- 

 gascar, there are numerous immersed corallites. These are closely arranged, and their walls 

 are well defined in longitudinal section, and well-developed trabeculae pass completely across 

 the polyp-cavities. Between neighbouring corallites the connecting structures are in 

 successive layers or storeys of laminae separated by rows of small, irregular, and short pillars. 

 On the upper surface between the immersed corallites numerous minute, broad-based, sharp- 

 pointed spinules occur, arising from a perforated calcareous lamina ; this is precisely the 

 condition of the successive layers below the surface, and new layers are evidently here added 

 in the same manner as on the corallite-wall. 



The third species described is referred by Duncan to M. granulosa, Ed. & H., but judging 

 from the description, and from small carmine-stained fragments which I believe to be the 

 material investigated, I am inclined to think it more nearly related to M. secale, Studer. In 

 any case the species serves to illustrate the structure of slow-growing varieties in which the 

 corallum is compact. The surface is here very dense, and consists of a stout lamina with 

 very few perforations, and is clothed with knobbed spinules in place of costae. In transverse 

 section concentric laminae, as in other cases, are separated by radially disposed bars which 

 represent former spinules. In this case, however, the laminae are much thicker and denser 

 and are closer together. Exceedingly narrow tubes lead from the polyp -cavities to the surface, 

 but their apertures are difficult to detect, as they are surrounded by the bases of spines. 

 Little need at present be added to the results of Duncan. Evidently the condition of the 



