91 . 



impossible that these represented another species. They are remarkable for 

 their small size, being rarely 10 cm. usually only a few centimeters high. The 

 caliculate or funnel-shaped body is, in accordance with its small height, very 

 narrow and its diverticula are very thin. The latter measure only 3-5 mm. in 

 thickness and are on an average 10-15 mm. long. It is also to be remarked 

 that in these specimens the cavities of adjacent calicos or funnels are united 

 much more extensively than in other species. Their radial, tubular diverticula 

 very frequently reach the next funnel, whereupon an open communication is 

 estabUshed between the cavity of the diverticulum and of the adjacent funnel 

 by the resorption of the intervening septum. As the radial, tubular diverticula 

 attain a width nearly equal to that of the funnels themselves, the whole 

 sponge appears as a network of tubes, in which the limits between the diverticula 

 (I'adial tubes) belonging to different funnels can hardly be made out. And 

 this limitation is rendered still more indistinct and difficult to determine by the 

 habit of the radial tubes terminally to divide dichotomously. This habit 

 of these specimens makes it quite possible to confound them, particularly when 

 they are merely isolated fragments, with Aphrocallistes ramosus. 



From Aphrocallistes heatrix Gray they can be ■ distinguished by the radial 

 tubular diverticula which do not, as in that species, increase in size towards 

 the base of the sponge, but which on the contrary become longer towards the 

 upper end of the sponge. This is a constant character, met with in all speci- 

 mens of Aphrocallistes bocagei hitherto described. 



The dictyonal net corresponds, as far as the most essential points of its 

 structure are concerned, with the descriptions given of the skeleton of Aphrocal- 

 listes bocagei Perc. Wright. 



The formation of the typical hexagonal cells of the honeycomb, open at 

 both ends, the irregular polyradiate nodes and the tuberculous, conical beams, 

 which protrude from it towards the dermal and gastral surface are quite the same. 

 In some points however, as the thickness of the septa and the triangularly prism- 

 atic pillars dividing the cells of the honeycomb, the thickness of the beams, 

 the sculpture and degree of distinctness of the nodes and particularly the number 

 and density of the small tubercles or spines covering the beams and nodes, such 

 great differences are met with, that I cannot consider these peculiarities as having 

 any systematic importance and I therefore do not believe that they can be used 

 as classificatory characters for the purpose of distinguishing the species. I 

 think it much more likely that they are merely individual characters, partly 

 perhaps dependent on the age of the specimen. It seems to me that at first, 

 that is to say in young specimens, the beams are slender and quite smooth, and 

 that they become thicker and more rough as the sponge grows older. Together 

 with the beams, the septa, and particularly the triangularly prismatic columns 

 dividing the cells of the honeycomb, increase in thickness and the nodular 



