272 DISCOVERY REPORTS 



case, the suspicion that the two species are identical is unavoidable. The present speci- 

 mens offer abundant evidence that this is in fact the case. 



The Discovery collection contains eight specimens ; five of these had been provision- 

 ally identified by me as Microxina charcoti and three as Gelliodes benedeni. On re- 

 examination, I found that several of the former contained isolated long, slender raphides, 

 and this was also the case in each specimen of the second species, but they were never 

 grouped in dragmata as in Gelliodes benedeni, var. fortior. Secondly, one specimen of 

 G. benedeni contained sigmata only, a second contained sigmata and an occasional 

 microxeote, and the third contained all transitions from sigmata to microxea (Fig. lo). 

 In addition, it is worth noting that the sigmata in Gelliodes benedeni are frequently 

 centrotylote, like the microxea of Microxifia charcoti. It is evident therefore that we 

 have to deal here with a single species, the microscleres of which may be sigmata and/or 

 microxea, to which raphides, either isolated or in dragmata, may be added. 



Distribution. Graham Land ; Victoria Land ; Wilhelm Land. 



Genus Hemigellius, gen.n. 



Genotype. Gellius nidis, Topsent. 



Diagnosis. Haploscleridae with skeleton composed of a coarse and irregular reticula- 

 tion of oxea formed by multispicular primary fibres, running vertically to surface, and 

 uni- or multispicular connecting fibres; special dermal skeleton formed by ends of 

 primary fibres spreading out beneath ectosome in a paniculate manner; microscleres, 

 when present, sigmata. 



Remarks. The original description gives little information regarding the structure of 

 the skeleton in Gellius riidis (see Topsent, 1901, p. 14) ; but the surface is said to be finely 

 hispid, which at least indicates the absence of a tangential dermal skeleton. Kirk- 

 patrick's (1908) specimens assigned to this species resemble the holotype sufficiently, 

 however, to justify the framing of a diagnosis, as has been done here, on the characters 

 observed in his specimen. 



The genus is closely allied to Haliclona. 



Hemigellius rudis (Topsent). 



Gellius rudis, Topsent, 1901, p. 6; id., 1901, p. 14, pi. i, fig. 9, pi. iii, fig. 4; id., 1907, p. 77, 

 pi. iii, fig. 2; G.fimbriatus, Kirkpatrick, 1907, p. 86; G. rudis, id., 1908, p. 45, pi. xvii, fig. i ; 

 G. fimbriatus, id., loc. cit., p. 46, pi. xvii, fig. 2, pi. xxiv, fig. 2; G. rudis, Hentschel, 1914, 

 p. 287; Burton, 1929, p. 422. 



Occurrence. St. 39: South Georgia, 179-235 m.; St. 42: South Georgia, 120-204 m.; St. 170: 

 Clarence Island, 342 m. 



Remarks. In a former work {loc. cit.) I included Gellius pilosus, Kirkpatrick, and 

 G. tenelltis, Topsent, as synonyms of this species. As shown on p. 276 the latter is 

 now included in the genus Adocia. As regards G. pilosus, Kirkpatrick, there can be 



