REVISION OF FRESHWATER SPONGES OF SPONGILLIDAE 63 



Discussion. — With the restoration of Gray's (1867) sj^stem, the 

 restriction of the genus Spongilla Lamarck, and the elevation of 

 Stratospongilla Annandale to fuU generic rank, it became necessary 

 to define a new genus, RadiospongUla, for the inclusion of those 

 members of the formerly recognized subfamily Spongillinae, which 

 possess gemmules with a thick pneumatic coat and a conspicuous porus 

 tube, more or less radially arranged gemmoscleres with a characteristic 

 arrangement of terminal spines, and lack true microscleres in their 

 dermal membrane and inner symplasm. All these species display a 

 number of constant intrageneric characters and form a distinct group 

 of intermediate position between the "Spongillinae" and "Meyeninae" 

 of previous authors, so that a further retention of these subfamilies 

 has become meaningless and futile. 



RadiospongUla consists of two clearly separable groups of species. 

 One is represented by R. cerebellata (Bowerbank), with which S. 

 biseriata Weltner, S. reticulata Annandale, S. proliferens Annandale, 

 S. semispongilla Annandale, S. micron Annandale, and S. sectospina 

 Rezvoj are synonymous, and is characterized by the possession of 

 generally smooth megascleres and a double layer of gemmoscleres. 

 The other consists of a large niunber of closely related species, dis- 

 playing spiny megascleres and a singular arrangement of gemmoscleres, 

 typically represented by R. sceptroides (HasweU) and R. craterijonnis 

 (Potts). R. cantonensis (Gee), elevated to full specific rank in this 

 paper, represents an extreme of this latter group, almost linking up to 

 spicular conditions found in Umborotula, but is still a clearly defined 

 species of the present genus. 



A number of species, which certainly also belong to this latter group 

 of RadiospongUla spp., must still be considered as insufficiently known. 

 Since most of them are not represented in the material available for 

 the present study, the assessment of their intrageneric relationships 

 is impossible. They are S. luzonensis Gee (1932) from the Philip- 

 pines, with which S. hozawai Sasaki (1936) from Japan most probably 

 is synonpnous, and two additional Japanese species recorded by 

 Sasaki (1936), S. sendai and E. craterijormw, the latter appears 

 closely related to R. indica (Annandale) . 



Four additional species recently found in Australian waters will 

 be described in a forthcoming paper (Racek, MS.), one of these already 

 recorded under the erroneous name S. botryoides Haswell by Gee 

 (1931d). In spite of the extensive search for this highly dubious species 

 of HasweU (1882) in the Australian region, and the great number of 

 specimens and species collected, no spongUlid has ever been found to 

 correspond, even vaguely, with the short description of the original 

 author. Since a type specimen of S. botryoides apparently has never 



