EEPORT ON CORALS — HYDROCORALLIN^E. 37 



apposed minute glistening white granules. The margins of the mouths of the dactylopores 

 are often slightly raised above the general surface. 



The older pores of the coenosteum are very deep and, as may be seen in longitudinal 

 sections of the branches or stems (t.o.), commence deep down within the stem near its 

 axis, and bend outwards on all sides to the surface of the branch with a nearly uniform 

 curve. The ccenostea of all Stylasteridae are traversed in all directions by a system of 

 freely anastomosing and branching canals. In the case of Sporadojiora, these canals are 

 especially abundant and form comparatively close meshworks, hence the whole coenosteum 

 is spongy and excessively porous when seen in section (PL II. fig. 1). The coenosteum may, 

 with most truth, be said to be built up of a series of hard partition walls, intervening 

 between and enclosing a highly complex system of tortuous canals and cavities. The 

 meshwork formed by these canals is closer and smaller towards the surface of the 

 coenosteum, more open and with wider meshes in the deeper regions. In the deeper 

 regions the main canals, as will be seen from the figure, follow more or less the 

 curved directions taken by the walls of the pores on their way towards the surface. 

 There is no main system of canals in the axis of the stem connecting all the zooid 

 cavities. The deep canals become more or less filled up, and the only connection 

 between distant zooids is by the more superficial living meshworks. In some places 

 irregular cavities of some extent occur amongst the smaller canals, and beneath the 

 ampulla (PI. II. fig. 1, G). At the very surface, the canal reticulation is very fine 

 indeed. 



The pores are cylindrical pits sunk in the spongy mass of the coenosteum, and their 

 walls are perforated all over by the openings of numerous canals. At their bottoms their 

 cavities pass off into a few large main canals of the meshwork. The styles of the 

 gastropores are very long, and can be traced deep into the axes of the branches of 

 coenosteum, they having become elongated as the growth of the pores and coenosteum required 

 it. In their deeper regions, these slender styles show a surface composed of a few 

 dentate ridges (PL XXXV. fig. 1, S) only, whilst in their upper and functionally active 

 region they terminate in a long brush-like mass, composed of complicated branchings of 

 fine and delicate calcareous spicules. At the base of this brush-like part of the style, a 

 very thin calcareous partition or " tabula" (PL II. fig. 1, T) is sometimes present, stretched 

 across the pore cavity at right angles to its axis. Sometimes two or three such tabulae 

 are present in a single gastropore, placed at successively deeper intervals. In some 

 instances, two tabulae occur close together in a pore, one above the other. These 

 tabulae are so excessively thin that I considered them at first to be membranous, but I 

 have been unable to dissolve them by the use of very strong alkalis, and I am now 

 convinced that they are calcareous. They do not seem to occur in all the gastropores, 

 and I have not observed them in any instance in the dactylopores. The dactylopores 

 vary much in size, as wdl be seen from the figures. 



