444 



THE VOYAGE OF H.M.S. CHALLENGER. 



iug to the form of the latter either a flat surface or one covered with numerous conical 

 elevations is formed. A similar and similarly perforated membrane is found also on the 

 opposed surface of the body-wall, the surface of egress, which indeed generally encloses an 

 internal gastric cavity, but may also, as in the case of many flattened or mushroom- 

 shaped sponges, be quite free and form an upper or lateral surface. 



" Between these two perforated boundary-surfaces there extends the simple strongly 

 folded layer of the ciliated cavities, which usually manifest a saccular shape, as I have 

 already described in Euplectella aspergillum, 1 but in some cases, as in the family Hyalo- 



nematidae, diverge to some extent from this. The 

 delicate wall of the cavities allows the square lattice- 

 marking to be perceived as in Euplectella, and is 

 also more or less thickly but irregularly perforated 

 by round pores. This system of ciliated cavities is 

 connected with the two boundary-surfaces by means 

 of a wide-meshed tissue of delicate anastomosing 

 trabecular, which are suspended and stretched between 

 them. Since, then, all the chambers are in direct 

 communication, and since their convex surfaces are 

 always turned towards the entering water, this latter 

 must flow through them in such a manner that it 

 enters through the pores and passes out through the 

 wide oral opening. 



" On account of the great uniformity in the struc- 

 ture of the soft parts, I have only been able to use 

 these for systematic purposes in a few cases, such as 

 in the definition of the Hyalonematidse. For such 

 purposes the form and arrangement of the siliceous 

 skeleton, which has hitherto been almost exclusively 

 applied by all spongiologists, is most significant. 

 " The two primary divisions of the Hexactinellida, Lyssacina and Dictyonina, which 

 Zittel founded some years ago in his important work on fossil sponges, I retain with 

 the same significance, but in consequence of my investigations I have been obliged to 

 modify his original definitions to some extent. 



" Zittel regards as Lyssacina those Hexactinellida in which the whole skeleton consists 

 of spicules which are only connected by means of the sarcode (exceptionally, however, 

 irregularly by means of flattened siliceous bodies), and in which the spicules of the soft 

 parts are for the most part very plentiful and highly differentiated. 



" The Dictyontna he defined as those Hexactinellida whose spicules are so united that 



1 Tram. Hoy. Hoc. Edin., vol. xxix. pp. 661-673, 1881. 



Fig. 166. — Melitlw.vlus ramosiis, n. gen. et sp., 

 a representative of the Uncinataria. 



