358 REVISION OF THE MONAXONID SPONGES, ii., 



(mostly spicule-) fragments, (ii.) of sparsely scattered subty- 

 lostyli measuring rarely as much as 130 by 3-5/x and (iii.) of 

 a very few, scattered, slender toxa and anisochelge, the latter 

 measuring, at most, 17/x long. It is quite a different type of 

 sponge, therefore, from that denoted by Lendenf eld's descrip- 

 tion, having no feature of resemblance thereto except an in- 

 crusting habit of growth, and even in this respect being not 

 quite similar, since the layer it forms is only about 1mm. in 

 thickness. Accordingly, in my opinion, it cannot possibly be 

 accepted as the type-specimen. 



A fragment from the British Museum, labelled StyJoteUa 

 aplysillioides, is also totally unlike the described sponge of 

 that name, and belongs to the genus Dendoricella — its spicules 

 being skeletal oxea, dermal tylota, isochelae arcuatae, and two 

 sizes of sigmata. 



Hence we are left with no clue as to the identity of Sfylo- 

 tella (rpJi/sinioides except its rather brief description, which, if 

 it can be relied upon, indicates that the correct position of the 

 species is in the genus Hymeniaculon. To this genus, then, 

 the species may, for the present, be regarded as belonging. 

 The only other species of Hymemacidon known from Port 

 Jackson, is that recorded by Ridley(33) under the name H. 

 caruiicuht Bowerbank ; this is also a horizontally extended 

 sponge with surface-elevations, but its spicules are stated to 

 attain a size of 290 by 8 ft, while those of H. aplyf^illioides, 

 according to Lendenfeld, measure only 130 by 6 /x. 



Subfamilia Phlceodictyin^. 



Rhizochalina ramsayi. (PI. XX., figs. 2 5; PI. xxi., fig.4; and 

 text-fig.8). 



The types consist of three half -specimens (derived by vertical 

 bisection of the originals), and a thick, median, vertical slice 

 of a fourth specimen. The sponge is massive, more or less glo- 

 bose, provided on its upper aspect with numerous thin-walled 

 erect fistulae, and below with few (sometimes only one) or 



