300 REVISION OF THE MONAXONID SPONGES,!., 



width. In reference to the pattern of this network, I need only 

 mention here that, in the peripheral layer of the sponge, to a 

 considerable distance below t.he surface, the trabecules are so 

 arranged as to form incomplete boundaries between elongated 

 " cells," the outer ends of which correspond in position with the 

 areolae of the surface, and the disposition of which, relatively to 

 one another and to the exterior, is exactly similar to that of the 

 cells of a honeycomb. In the case of the small specimen before 

 me, the trabecule forming these cells still retain their separate 

 individuality, thus enabling one clearly to distinguish between 

 (i.)main ones, relatively few in number, running in the longi- 

 tudinal direction of the cells, i.e., perpendicularly to the surface, 

 and (ii.)more numerous transverse or connecting ones; but in 

 the large (and older) specimens, presumably as the result of the 

 increase in width and gradual concrescence of the trabeculse, and 

 of the consequent reduction (even to the point of complete 

 obliteration) of the intervening meshes, the condition is such 

 that the cells are divided from one another by almost or quite 

 comp'ete partitions, and thus bear a structural likeness to the 

 cells of honeycomb, which is almost perfect. 



The cortical rind, which is of very firm, dense, and fairly tough 

 consistency, varies in thickness, in the different parts of the 

 surface, from about 0-5 mm. to upwards of 1 mm. In the macer- 

 ated sponge, it separates from the underlying skeleton with the 

 greatest ease, and is then seen to be not less thick, or scarcely 

 less thick, at the position of the areolae than elsewhere; accord- 

 ingly, the original description seems again to be at variance with 

 fact, when it speaks of " membranes which extend in the meshes 

 of the surface-network," and mentions, further, that these mem- 

 branes have " groups of smaU pores " situated in them and are 

 " very thin and delicate." The skeleton of the cortex, apart 

 from a thin external layer of microscleres (of the two non-oxeote 

 kinds) consists of closely packed tylostyles, the most superficially 

 situated of which are disposed vertically to the surface; within 

 the circular meshes or areolae, the skeleton is much less dense, 

 and the cortex is, consequently, much softer than in the intervals 

 between. 



