DENDY— REPORT ON THE HEXACTINELLID SPONGES (TRIAXONIDA) 215 



thus constituted is roughened with small conical spines, but these are more strongly 

 developed on the subdermal trabeculse than deeper down. 



At various points rather small, spiny-rayed hexacts are seen to be undergoing 

 incorporation in the general framework by fusion of their rays with the trabeculse. To 

 what degree the framework really grows by this incorporation of originally separate hexacts 

 it is impossible to say, but the following observations rather suggest that it extends 

 centripetally in this manner. 



For some little distance beneath the gastral surface, in the region occupied by the 

 inner trabecular layer of soft tissues, the main dictyonal framework is absent and its place 

 is taken by scattered hexacts quite irregularly arranged. These hexacts (figs. 12, 17, hex.) 

 are far more numerous here than anywhere else in the sponge, and it is at the junction of 

 this layer with the chamber-bearing layer that the incorporation of hexacts in the dictyonal 

 framework is chiefly seen (figs. 12, 18). 



Possibly growth of the main skeleton also takes place by the formation and subsequent 

 fusion of outgrowths from the trabecular themselves, as in Sarostegia oculata, but it is often 

 diflficult to distinguish between such outgrowths and the projecting rays of partially 

 incorporated hexacts. 



The hexacts of the subgastral layer also frequently unite with one another by fusion 

 of rays, and thus tend to form a very irregular dictyonal framework with much smaller 

 meshes and more slender trabeculse than those of the main skeleton (fig. 12, hex.). 



Sometimes even discohexasters may be incorporated in the skeletal framework, 

 giving rise to very curious appearances (fig. 13). 



The parenchymal spicules are as follows : — 



(1) Spiny-rayed hexacts (figs. 12, hex., 14); with rays straight or slightly curved, 

 and varying a good deal in length and thickness ; sometimes sharply pointed at the ends 

 and sometimes more or less clubbed. These spicules occur chiefly in the subgastral layer, 

 where they are united together by fusion of rays into an irregular, loose network, while 

 the outer ones are also united in the same manner with the inner portion of the main 

 dictyonal framework. It is difiicult to find a single hexact lying entirely free in the 

 parenchyma. 



(2) Oxyhexasters (fig. 15) ; with long, slender, sharp-pointed rays. These appear to 

 be extremely rare and are perhaps not a normal constituent of the spiculation. 



(3) Discohexasters (fig. 16) ; with slender, curved rays terminating in toothed discs 

 and varying much in length in different specimens. There are usually about five terminal 

 rays to each principal. These spicules are very abundant in the subgastral portion of the 

 sponge wall. Occasionally they become incorporated in the skeletal framework (fig. 13). 



This sponge is sufficiently well preserved to enable me to give some account of the 

 structure of the soft parts (figs. 17, 18). A very delicate dermal membrane (d.m.) is 

 stretched over the outer surface of the main skeletal framework. This framework is 

 interrupted at frequent intervals by the rather wide inhalant canals (i.e.). The dermal 

 membrane' still extends over the outer ends of these canals in some cases as a thin net 

 pierced by the inhalant pores, but for the most part it is absent from the openings in 

 question, perhaps owing to abrasion or shrinkage. Beneath the dermal membrane comes 



