220 PERCY SLADEN TRUST EXPEDITION 



of sarulee. These collars, however, are only preserved where the surface of the sponge 

 has been protected from rubbing, as shown in fig. 19. In one case (fig. 19/) two polyps, 

 apparently formed by fission of a single one, were observed within the same collar. The 

 polyps appear to be connected with one another by a network of stolons ramifying in the 

 thickness of the sponge wall. 



Similar polyps were described by Topsent in his specimen, and Schulze [1904] speaks 

 of the macerated fragments obtained by the " Valdivia " as showing " dellenartige Ober- 

 flachenvertiefungen von ovaler oder doch rundlicher Form mit schwach erhabenem 

 Rande." It seems probable that these are the shallow depressions left after the removal 

 of the polyps. The association between sponge and polyp thus appears to be a 

 constant one. 



The main skeleton is a close framework of usually stout trabeculse (up to about 

 0'07 mm. in thickness), with triangular meshes. As many as six bars of this frame- 

 work may radiate from a common centre in approximately the same plane*, like the 

 spokes of a wheel, connected at their outer ends by the spokes of similar adjacent 

 systems, giving the whole framework a very characteristic appeax-ance, as shown in 

 fig. 33. 



Apparently the whole framework grows not so much by incorporation of new 

 hexacts as by the outgrowth of secondary trabeculse, whose ends meet and fuse to form 

 systems similar to those just described. This process is certainly responsible for the 

 inward extension of the framework by which the original central cavity becomes more 

 or less obliterated. 



The structure of the dictyonal framework thus agrees very closely with Schulze's 

 description and figure of Ramella tuhulosa, except that what I may perhaps term the 

 " rotulate " character, due to the formation of triangular meshes, appears to be more 

 strongly pronounced (fig. 33). I have no doubt, however, that this character is a very 

 variable one. In the older parts of the skeleton the trabeculas are smooth, but in the 

 younger parts, adjacent to the dermal and gastral surfaces, where the trabecules are 

 more slender, they are often roughened with minute projections (fig. 34). 



The manner in which the dictyonal framework spreads into the central cavity is 

 well shown in figs. 35 and 36. Slender, moi^e or less radially arranged outgrowths 

 are given off from the superficial trabeculse of the gastral surface (fig. S5, pr.), and their 

 ends, coming in contact with one another, fuse to form a new node of the skeleton 

 (fig. 36, pr.). Doubtless these slender processes, which are at first minutely roughened, 

 are thickened and become smooth later on by the addition of concentric layers of silica. 

 Growth seems to take place at the dermal surface of the dictyonal framework in 

 precisely the same manner, and probably some, at any rate, of the freely projecting, 

 minutely spiny knobs, which occur on this surface, are the immediate agents concerned 

 therein. I have seen just the same fusion of such outgrowths to form a new node on the 

 dermal as on the gastral surface. Some of the minutely spiny knobs on the dermal 

 surface, however, appear to be the reduced centrifugal rays of the outermost fused hexacts 

 of the original framework, as described by Topsent. 



* Of course similar bars radiate from the same centre in other planes. 



