388 DISCOVERY REPORTS 



from 22-42 per cent: the funnel-organ is double, its components are thick and short. 

 The hectocotylus is 8-1 1 per cent of the third arm. The sculpture is never as close and 

 as well developed as that of E. charcoti, though it is of the same type. I regard this form 

 as conspecific with E. charcoti. 



(2) A broad-bodied, narrow-headed form with exceptionally short arms (57 per cent) 

 and very deep web (60 per cent). The funnel-organ is double. The gill-filaments are 

 exceptionally low (five to six) in number. The eyes are large, but not adnate. The sculp- 

 ture is highly characteristic, being composed of flattish irregular, very clear-cut, 

 tubercles, rather like those of O.pallidus. It has no ink sac and the radula is degenerate. 

 It is so distinctive that I must treat it not only as a new species, but as referable to a 

 new genus (p. 392). 



(3) A narrower, smooth or obscurely granular form, with very large eyes. The head 

 is not much narrower than the body. The hectocotylus is rather small (6-8). The funnel- 

 organ is double. This I regard as Joubin's E. turqueti, though the original specimen of 

 the latter was very small and, at the same time, unlike those figured subsequently by 

 Joubin. I rely on the later figures for my identification. 



(4) More uncertainty invests the fourth form, which seems to occur in two phases, 

 one of which seems to be related to certain forms oiE. charcoti. I provisionally treat this 

 as a distinct species (E. polymorpha). The skin is granular, the eyes moderate to large, 

 the funnel-organ is single and the hectocotylus large (17-9 per cent). It appears in two 

 phases: (a) a narrow form with the head as wide as the body and a deep web, and (b) 

 a more obese form with a shallower web. This form has a marked tendency to bear a 

 peripheral keel. In other respects it is not unlike E. charcoti. 



I have carefully considered the relationship of the new species with those given in 

 Berry's key (1917, p. 14), and can find no likeness to any in the latter. E. polymorpha 

 has certain features in common with Berry's E. harrissoni. This matter is discussed on 

 p. 392 and the possibility that these forms may be conspecific is set aside as most unlikely. 



Consideration of the structure of the radula, hectocotylus and other organs leads me 

 to believe that this genus should be sub-divided into several sub-genera to render 

 apparent certain marked structural divergences. Some of these have been already 

 proposed, but I defer to a later publication the complete rearrangement of the group. 

 For a preliminary discussion and remarks on the interesting new genus, Thaiimeledone , 

 see pp. 374, 392. 



Genus Graneledone, Joubin 



Graneledone charcoti, Joubin. 



Eledone Charcoti, Joubin (1905, p. 22). 



Thirteen examples (five (?<?, eight 9?) from South Georgia and vicinity. 



St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. 179-235 m., grey mud. Large otter 

 trawl: three ?? (Brit. Mus. 10 (i-iii)). 



St. 42. i.iv. 26. Mouth of Cumberland Bay, South Georgia. 120-204 m., mud. Large otter 

 trawl: one S (Brit. Mus. 12). 



