Bd. V: 6) THE FISHES OF THE SWEDISH SOUTH POLAR EXPEDITION. 35 



I specimen caught on long-lines in a depth of lOO m., clayey bottom, Cumber- 

 land Bay, South Georgia. 20th of May 1902. 



1 specimen from stat. 22, Cumberland Bay, South Georgia, depth 75 m., clay 

 and some algœ, temperature at the bottom -f- 1,5° C. 14th of May 1902. 



1 specimen from stat. 32, South Fjord, Cumberland Bay, South Georgia, depth 

 195 m., clay mixed with stones, temperature at the bottom + 1,45' C. 29th of 

 May 1902. 



2 young specimens from stat. 2i3i Boiler Harbour, Cumberland Bay, South Geor- 

 gia, depth 22 m., clay and algae. 30th of May 1902. 



In his repeatedly quoted paper on the fishes of the -'Southern Cross" Expedi- 

 tion BOULENGER puts N. Diarrnorata FISCHER as fully synonymous with N. macro- 

 cephala GÜNTHER. Although there cannot prevail any doubt about the close re- 

 lationship between the two forms which have received the name just mentioned I 

 think it is most correct to maintain FisCHER's name marmorata as a mark of sub- 

 specific distinction for the fish of South Georgia. There appears namely to be some 

 constant differences between the latter and those from other localities. If firstly the 

 number of fin-rays is considered, the fish from South Georgia seems to differ in a 

 remarkable degree. BouLENGER indicates (1. c.) the formula of the dorsal to be 

 IV, 29 — 30, and that of the anal 23 — 25 in the typical N. viacrocepJiala. The ori- 

 ginal diagnose of GÜNTHER* had "D. V, 30—31, A. 21." In one specimen out of 

 fifteen I have found only 4 spines in first dorsal, in four 5, in eight 6, and in two 7. 

 The rays of second dorsal were in two speciinens 33, in six 34, in six 35 and in one 

 37. The anal rays were in one specimen 27, in ten 28 and in four 29. The formula 

 for the fish from South Georgia may be compiled from this to be D. (IV) V — VII, 

 33—35 (37); A. (27) 28 — 29. A racial or subspecific difference appears to be clearly 

 founded on these facts. The number of scales as well, is greater in the subspecies 

 viannorata in which a longitudinal series of scales above the lateral line appears to 

 contain from 6j to '/6 scales, against 58 to 62 in the typical macrocepliala according 

 to BoULENGER (1. c). FiSCHER based his diagnose of N. uiarvwrata only on three 

 specimens which appear to have had a, for this subspecies comparatively low number 

 of rays, so that the formula in the original description of N. marmorata was D. V, 

 33; A. 26 (28). The small number of type specimens together with this less ab- 

 errant formula for the fin rays probably made BoULENGER assume the identity of 

 macrocephala and marmorata. 



The interorbital width appears to be smaller in A', marmorata than in the ty- 

 pical A'^ macrocephala. The latter is namely said to have its interorbital width con- 

 tained only 2V3 to 2V2 times in length of head, while I have found the same rela- 



* Cat. Fishes Brit. Mus. Vol. II. London i860, p. 263. 



