SILICIOUS AND HORNY SPONGES WILSON. 449 



joined " together to form irregular sinuous canals, having a some- 

 what stellate arrangement," which are " seen as dark spaces through 

 the skin" (Sollas, 1888), is not constant. The species remains 

 characterized especially by the skeletal elements, in particular by the 

 phyllotriaenes. These spicules are arranged in several layers in the 

 ectosome; the rhabdome is short; the clads long and narrow with 

 irregular outline, very variable in the same specimen, and indeed 

 in the same spicule, sometimes simple, sometimes irregularly bifur- 

 cate or even trifurcate — that is, with secondary branches that divide. 



One of the specimens from D5218 is about cylindrical, tapering 

 toward the upper end, passing below into an expanded amorphous 

 basal part; 65 mm. high, 35 mm. in transverse diameter. An os- 

 culum, 6 mm. in diameter, at the upper end passes into a cylindrical 

 cloaca which extends down nearly to the base of the sponge. The 

 second specimen, 80 mm. high with transverse diameter of 50 mm., 

 from this locality is very similar. The base of this specimen has 

 been torn off. The osculum measures 10 mm., the cloaca 15 mm., 

 in cross diameter. Color of these sponges, a light brown. 



The sponge from D5593 is 90 mm. in height, with a transverse 

 diameter of 50 mm. in the upper half, widening below; vase-like; 

 osculum 15 mm. in diameter; cloaca 18 mm. in diameter, tapering 

 below, 45 mm. deep. The apertures of the very numerous canals 

 opening into it are not all open, some being closed in with sieve- 

 membranes. 



The specimen from D5252 is not vasif orm, but massive ; somewhat 

 elongated vertically, larger above, smaller below near the basal sur- 

 face by which sponge was attached; height 95 mm., horizontal di- 

 ameter 80 mm. above, 50 mm. below. The somewhat flattened upper 

 surface bears four oscula, 8-10 mm. in diameter, each leading into a 

 cloacal canal of about the same width and 20^30 mm. deep. 



In none of these specimens does the surface bear marked ridges 

 or protuberances. 



The pores are strictly not in groups, although they may appear so. 

 The fact is that the dermal membrane, riddled everywhere with 

 pores, rests immediately on narrow subdermal trabeculae about 60 

 mm. thick, which include the clads of the phyllotriaenes and which 

 surround and inclose small subdermal chambers, 150-230 (jl in di- 

 ameter. Into each of these chambers a few, 1-5, pores open. Balsam 

 preparations show that, as is so often the case in sponges, there are 

 pores directly over the subdermal trabeculae as well as between them, 

 the trabeculae only appearing to be solid. 



The larger subdermal spaces, other than the small uniformly dis- 

 tributed ones just referred to, as seen through the dermal membrane, 

 vary in appearance in the different specimens. In those from D5218 



