HEXACTINELLIDA 2SS 



completely intermediate between the two older varieties in these three distinguishing 

 features : the sponge may have a basal tuft or may be attached directly to the substratum, 

 the velum is moderately well developed (in one specimen it is present on one side of the 

 sponge and completely absent on the other, although the diactin pleuralia are equally well 

 developed all over), and the rays of the pentactin pleuralia are confined usually to a 

 sector of i8o'. 



With regard to the question whether a species which normally has a basal tuft may be 

 represented by specimens attached directly to the substratum (cf. Kirkpatrick, 1907, 

 p. 6), the present specimens contribute some relevant data. The individuals of var. 

 intermedia seem to be able equally well to produce a basal tuft or to attach themselves 

 direct, and there is evidence that the same thing is true, though perhaps to a lesser 

 extent, in R. niida, Topsent, and R. racovitzae, Topsent {q.v.). 



One of the specimens is a " twin", recalling the specimen of var. solida originally de- 

 scribed by Schulze (1887) from Kerguelen. 



Among the specimens from St. 156 is a small bundle of pentactin pleuralia from an 

 individual of var. intermedia bearing a bud about 3 mm. in length. The presence of this 

 bud in a jar which included, together with non-Hexactinellid sponges, only a small 

 specimen of the present variety, suggests that it was found free and had not been 

 accidentally torn away from a larger, budding specimen. If correct, this surmise offers 

 an explanation as to how the buds of R. antarctica, which migrate along the shafts of the 

 pentactin pleuralia, get over the obstacle offered by the rays at the ends of the pleuralia. 

 Presumably the pentacts bearing the buds are shed and the buds thus released, 



Rossella nuda, Topsent (Figs, i, 2). 



(For synonymy see Burton, 1929, p. 409.) 



Occurrence. St. 39: South Georgia, 179-235 m-l St. 45= South Georgia, 238-27001.; St. 123: 

 South Georgia, 230-25001.; St. 148: South Georgia, 132-14801.; St. 190: Palmer Archipelago, 

 93-130 m.; St. WS 239: Falkland Islands, 196-193 m.; St. MS 68: South Georgia, 220-247 m. 



Remarks. The several specimens, only three of which are dry, range in size from a 

 few cm. to 77 cm. high, and show all the variations illustrated by me (1929, p. 406, 

 text-fig. i), including that of the hodgsoni form. Of the specimens from St. 190, one is 

 fragmentary and the other is a "twin" (cf. R. antarctica, var. intertnedia) 14 cm. high 

 and very like the holotype of " Aidorossella levis ". The specimens from Sts. 39 and MS 68 

 consist of papillae torn from the surfaces of large specimens of the c-iorm} The specimen 

 from St. 148 is sub-spherical, 10 cm. high, with surface almost apapillose, but with 

 strongly developed bundles of pleuralia projecting from the surface. There are only 

 vestiges of a root-tuft and the base is truncate with small pebbles, etc., buried in its outer 

 tissues. This specimen combines the sub-spherical shape of both the «-form and the 

 c-form, the surface of the i-form, and, in having no root-tuft, it approximates also to both 

 the e-Lrm and the /-form. The specimen from St. 123 is cyHndrical, 14 cm. high and 

 5 cm. in diameter, with slightly papillose surface beset with feeble bundles of pleuralia. 



1 These letters correspond to the letters given in my Terra Nova Report, 1929, text-fig. 2. 



3-2 



