128 ART. 7 1. IJIMA : HEXACTINELLTDA, IV. 



Three genera make up the subfamily. They can be distin- 

 guished from one another by the cliaracters indicated in the fol- 

 lowing key. 



<'. — Witliout pentactinic liypodermalia Acanlhascm F. E. ScH. 



''. — With peiatactinic liypodermalia. 



a'. The paratangentials of hypodermal pentaotins never pronged 



StaurocalyptuH I J, 



6". — Tlie same, when fully developed, armed with spine-like or hook-like prongs 



Rhabdiiccih/ptus F. E. h^CH. 



A fourtli genus, Acanthosaccus (witli tlie single species A 

 tenuis), was instituted by F. E. Schulz i-: ('99) on the ground 

 that it is totally lacking in oxyhexasters, instead of which there 

 should occur numerous minute oxyhexactins. The oxyhexactins 

 here referred to are in my opinion identical with the spicules called 

 by me the hexactinose oxyhexaster. So that, in my system of 

 terminology, Acanthosaccus F. E. Sch. is as really in possession 

 of oxyhexasters as any other genus of the subfamily. Whatever 

 be the name given to these spicules I am wholly skeptical as 

 to the expediency of regarding this sponge as generically distinct 

 from Rhahdocalyptus. 



As regards the organization of the subfamily a great deal 

 can be said in general. 



The AcanthasciucT are all moderately thick-walled, j^oucli- 

 like, barrel-like, cup-like, vase-like, or tubular forms with a deep 

 gastral cavity. The osculum, surrounded by a thin marginal 

 edge, occupies tlie superior end, while the inferior blind end, 

 which may be contracted into a short stalk-like l)ase, is as a 

 rule firmly attached to the sul)stratum. The only exception is 

 offered l^y R. jjlumodigllatus R. Kirkp. ('01), which is rooted in 

 the loose bottom l)y means of basal tufts of anchoring spicules. 



