THE SPONGES. 79 



to wliicli connectives have developed. Tlie plates are always set edgewise 

 to the surface, and in any particular part of the sponge body they nmay 

 be seen, in the macerated skeleton, to extend in a very general way in 

 one direction, which in some cases obviously corresponds to the long axis 

 of the lobe. Owing to the labyrinthine habitus of the sponge body, it 

 is impossible in other cases to determine which is the long axis of a 

 particular part. 



The arrangement of the constituent beams of the skeletal plate has a 

 certain regularity. Beams directed radially to the surface of the sponge 

 may be distinguished, between which lie connectives. The latter are 

 frequently transverse, thus giving rise to rectangular meshes, Figs. 1 and 

 7, Plate 11. The superficial ends of the radial beams form tapering 

 spines of varying length, sometimes very short, frequently as long as 

 350 fi, often slightly irregular, and as a rule thickly covered with sharp 

 microtubercles. The beams in general are sparsely covered with similar 

 tubercles, and most commonly have a thickness of 20-30 /a. A similar 

 difference in tuberculation between the ends of the radial beams and the 

 general framework exists in Hexactinella grimaldu Tops. (Topsent, 1892, 

 p. 34), and according to Topsent (1892) in H. hihulosa F. E. Sch. 



From the nodes of the skeletal reticulum, on the free surfaces and edges 

 of the plates, slender, sharp, tuberculated spines very generally project, 

 Fig. 1, Plate 11. Some of the very delicate connecting bars which extend 

 between adjoining skeletal plates (lower left corner and at extreme right, 

 Fig. 1, Plate 11) make the impression of having arisen through the fusion 

 of such spines. 



Commonly the outermost tangential beams of the dictyonal framework 

 lie at some considerable distance below the dermal membrane (Fig. 7, Plate 

 11). This is not always so, for in places they nearly, and more rarely 

 quite, reach the surface, aiding in support of the dermal membrane. The 

 lining membrane of the cloaca-like main canals is likewise in places 

 directly supported by beams of the framework, placed tangentially to 

 this membrane. The attached surface of the sponge lobes is compara- 

 tively smooth, the dictyonal framework here forming the familiar close 

 meshwork or "cribellate plate" (Ijima) found in so many Hexactinellid 

 sponges. 



In spots at the surface the dictyonal framework is covered with 

 collections of delicate hexacts and pentacts, partially fused with one an- 



