COLLECTED BY THE SCOTTISH NATIONAL ANTARCTIC EXPEDITION. 435 



The skeleton consists internally of a dense reticulation of closely packed 

 tylostyli. At the surface, the tylostyli, which are here shorter, are arranged in 

 vertical brushes, and their points project slightly beyond the dermis. The 

 spinispirse are not present in very great abundance, and do not form a dermal crust. 



The tylostyli have slightly curved shafts and well-rounded heads. They vary 

 in length between 0'2 mm. and 0*43 mm., and have a maximum diameter of 

 0'015 mm. In shape they are very similar to the stouter tylostyli of the 

 specimen of Hymeniacidon angulata Bowerbank, numbered B. M. 21 by Vosmaer 

 (27), and figured by him on pi. xiii, but they are, on the whole, fairly thick, 

 and their diameter does not often fall below O'Ol mm. The old specimen referred 

 to possesses a much greater proportion of more slender megascleres. 



A good many abnormal tylostyli are present, similar to those described and 

 figured by Vosmaer (27, p. 58, pi. xiv). 



The spinispiree are all slender, and are similar to those figured by Vosmaer 

 (27, pi. xiii, fig. l). The loDgest measured reached a length of 0"032 mm. As 

 already mentioned, they do not form a superficial crust. 



In attempting to assign to the Scotia specimens a place in one of the seven tropi 

 or groups into which Vosmaer divides the specimens of Spirastrella purpurea 

 examined by him, only the spiculation can be taken into account, as their external 

 form is unknown. The specimens placed in tropus tegens and in tropus tuberosa 

 have a glabrous surface and possess robust spinispirse, therefore the Scotia sponges 

 do not belong to either of these groups. Nor do they belong to the pyramidalis 

 group, the members of which also have a glabrous surface and often possess robust 

 spinispirse. Again, the Scotia specimens do not belong to the digitata group, 

 as the specimens placed in it have a superficial crust of spinispirse. They agree 

 with the specimens placed in the remaining three tropi, tubulifera, concrescens, 

 and glasbosa, in the absence of a superficial crust of microscleres, and in possessing 

 a slightly hispid surface. As these three groups are characterised by the external 

 appearance of the specimens placed in them, it is impossible to decide to which 

 of them the Scotia specimens most strictly belong. This is, doubtless, the less 

 to be regretted as Professor Vosmaer has shown so clearly the wonderful variations 

 of this species and the way in which one tropus passes into another, so that, 

 as he states, some specimens might be placed in one tropus equally as well 

 as in another. 



Vosmaer (27, p. 35) gives in full the geographical distribution of the species. 

 As regards the distribution in the Atlantic Ocean, the sponge has been taken 

 off the north coast of France and off Madeira. It has also been taken in the 

 Mediterranean. The finding of specimens off the western coast of South Africa 

 thus extends southwards the known distribution of the species in the Atlantic. 

 Other specimens of the species have been taken off South Africa, but they were 

 taken off the south-eastern and eastern coasts, namely, at Port Elizabeth (S. capensis 



