122 Prof. W. Thomson on the " Vitreous " Sponges. 



many cases several are partially united together. In Euplec- 

 tella perfect and distorted spicules of this class are entangled 

 in the loose meshes of a framework which is evidently formed 

 by the still fm-ther distortion and anastomosis of spicules of 

 the same type. In Aphrocallistes (see woodcut) the network 

 is still more evidently produced by the coalescence of stellate 

 spicules, though their hexradiate character is somewhat ob- 

 scured. In the network of Farrea (Bowerbank, Brit. Spong. 

 vol. i. fig. 277) the hexradiate type is very marked. The 

 primary axis of the spicule is reduced to a conical tuberculated 

 spine ; and spines of exactly the same form are developed in a 

 corresponding position on the outer surface of AjjhrocalUstes 

 (woodcut, e). In Farrea the spicules are distinctly tubular; 

 but this is merely a question of degree. In Dactylocalyx 

 the reticulation has become very irregular; but I have no 

 doubt, from the style of netting (see Bowerb. Brit. Spong. 

 vol. i. fig. 275) and from the close analogy in other respects 

 between this genus and Ajjhrocallistes and Farrea^ that its 

 fundamental plan is the same. 



I believe that it would be safe to accept the generalization 

 that the continuous siliceous network, wherever it occurs in 

 the vitreous Sponges, is produced by the fusion of spicules of 

 the hexradiate type. 



Throughout the whole order the spicules of the sarcode are 

 very abundant, and are often very elegant in form. They are 

 extremely minute, usually not more than '001 of an inch in 

 length or diameter, and are seen adhering in groups to the 

 larger spicules, or entangled in multitudes among the fascicles 

 and in the meshes of the sponge. However complex these 

 spicules may finally become, they all, with perhaps one doubt- 

 ful exception — the bihamate spicule figured by Bowerbank as 

 from Farrea occa (Brit. Spong. vol. i. fig. 114) — may evidently 

 be referred to the hexradiate type. In both species of Habro- 

 dictyon the most common sarcode-spicule is the " floricomo- 

 hexradiate " of Bowerbank (PI. IV. fig. 1 e) ; and the same 

 form occurs abundantly in Euplectella aspergillum. The 

 " coronato-hexradiate stellate " form figured by Bowerbank 

 (Brit. Spong. vol. i. fig. 195) seems to be the central star of 

 one of these without the cm-ved processes. Associated with 

 this type, we have in Habrodictyon multitudes of extremely 

 minute hexradiate spicules, variously armed and feathered, 

 and scarcely distinguishable from the ordinary spicules of the 

 sarcode of Hyalonema. The spicule which in the series de- 

 parts most from the hexradiate form is the wonderful double 

 grapnel of Hyalonema ; but, although its ends are split up into 

 curving flukes, in the very middle of the shaft the cross-canals 



