THE SPONGES. 109 



and longer axis, the spicule representing a fairly typical metaster. In a 

 representative spicule of this kind the axis was strongly curved ; the rays, 

 nine in number, 6-8 fi long. 



Poecillastra crihraria stands close to P. schidzei, SoUas (1888, p. 79), the 

 type specimens of which I have examined. The arrangement of pores and 

 oscula is very similar in the two forms. 



Note. The genus Poecillastra was created by Sollas (1888, p. 79) for 

 certain plate-like sponges collected by the '• Challenger." The preserva- 

 tion of these sponges was so good that Sollas was able to study their 

 anatomy, and in consequence to define his genus in an unusually satis- 

 factory manner. The characteristics of the new forms {P. schidzei, P. 

 crassiuscula, P. laminuris, P. tenuilaminaris) were the following : The body 

 is plate-like. On one surface are the pores, either evenly distributed or 

 in closely set areas. On the other surface are the oscula, small and evenly 

 dispersed. The main afferent and efferent canals are small, very numerous, 

 and their openings are thickly distributed over the corresponding surfaces" 

 Specialized pore areas, as in Thcnca, and specialized oscular areas, as in 

 Sphindrella, are not present. The flagellated chambers are eurypylous. 

 The megascleres are oxeas and tetraxons, the latter represented by ortho- 

 triaenes and calthrops, the two forms being very similar. The triaenes 

 are localized at the surface, the calthrops scattered through the interior. 

 Both triaenes and calthrops may he ''■comparatively rare" (P. crassiuscida). In 

 the triaenes and calthrops the actines may be "frequoiily rounded off into 

 short rods or tubercles" {P. temdluminans). The microscleres are microxeas, 

 spirasters, and metasters varj'ing to plesiasters'. 



The similarity of the two species described in this report to the four 

 " Challenger " species of Poecillastra is very close, and it is obviously 

 necessary to include them in the same genus, if we take into considera- 

 tion the entire structure. In respect to a single point, viz. the degener- 

 ation of the rhabdome, Poecillastra tricornis resembles the sponges grouped 

 by Sollas (1888) and Topsent (1902) under Nethea. This genus has been 

 a very doubtful one. Sollas (1888) erected it for Nethea {Tisiphonia) nana 

 Carter (1880). Topsent (1902) adds to it Nethea {Pachastrclla) amycj- 

 daloides Carter (1876) and Nethea (Pachastrella) connectens 0. Schm. 1870. 

 In all three sponges the canal system is unknown, and in the last the 

 skeleton is imperfectly known. It is possible therefore that the group is 

 heterogeneous. The new definition given by Topsent (1902, pp. 10-11), 



