346 DISCOVERY REPORTS 



This feature is best demonstrated by the 'first-slope' dwellers. Thus while A^. canina and N.jordani 

 inhabit much the same area off the mainland coast in the southern region, the latter was consistently 

 found in slightly deeper water than the former. A^. tessellata, with a depth distribution overlapping 

 that of both the last-named species, was not found along the same coasts, but mostly in more exposed 

 positions where the 'first slope' was steeper (e.g. round the Falkland Islands, where the other two 

 species did not occur at all). A^. longipes, if it is indeed a distinct species, seems limited to the western 

 end of the Magellan Channels, while the others were found to the eastward, mostly along open coasts. 

 Lastly, Harpagifer bispinis might almost have been placed in group III, for it is common intertidally 

 at times, but we have taken it down to 95 m. in the Patagonian region ; and in the Antarctic Zone, 

 where it has a far more extensive distribution regionally, it would certainly be more accurately 

 described as a ' first-slope ' dweller. 



The deep-water group consists mainly of larger fish, so that, partly owing to increased powers of 

 movement, there is considerably more overlapping between the species. Considering the Notothenia 

 spp. first, the two smallest, A'^. elegans and N. giintheri, seem to move but little and to remain on the 

 plain of the shelf throughout the year. A^. giintheri has its centre of distribution well to the south of 

 that of A'^. elegans. N. ramsayi, dominant throughout and the largest of the genus in the region, was 

 found to have a well-marked seasonal migration to deeper water beyond the shelf edge in winter. 

 Dissostichus eleginoides was too rare to be studied on these lines. It is the largest by far of all the sub- 

 Antarctic Nototheniiformes. Our large specimens were captured beyond the edge in very deep water, 

 but smaller ones were taken on the shelf, so it may move inshore to spawn. Cottoperca gobio and 

 Champsocephalus esox were mainly inhabitants of the plain of the shelf in the southern region, showing 

 considerable overlap with Notothenia ramsayi, but although a few Cottoperca were taken beyond the 

 shelf edge, neither of them seemed to show anything like the definite movement to deeper water of 

 Notothenia ramsayi. Moreover, both Cottoperca and Champsocephalus were frequently taken in lesser 

 depths than those to which Notothenia ramsayi normally penetrates— sometimes, indeed, in littoral 

 waters. 



The exclusively littoral group naturally show almost complete territorial overlapping, but their size 

 differences may serve to lessen the competition among them. A^. brevicaiida, N. squamiceps, N. sima 

 and A^. cornucola are all very small species and no doubt compete for small invertebrate food ; but 

 A^. macrocephala and A^. wiltoni run to a fair size (30 cm. or more) as Patagonian Nototheniidae go! and 

 here, doubtless, ' the great ones eat up the lesser ones '. Eleginops maclovinus may run up to 5 lb. weight, 

 but its diet is unknown. It would be extremely interesting to see whether the convergent evolution 

 evident in its close superficial resemblance to true mullets, and emulation of their powers of ascent 

 mto shallows and fresh water, extends also to the adoption of a vegetarian diet. 



In the matter of size generally, the 'first-slope' dwellers provide a majority of the intermediates 

 between the extremes shown by the smallest littoral species and the deep-water group. Thus, with 

 certain obvious exceptions, the division by depth distribution is broadly reflected in a corresponding 

 gradation in size, just as we have so often found within the limits of individual species. 



It is not yet possible to say much concerning the growth of Patagonian Nototheniiformes. Length 

 frequencies of Cottoperca gobio suggest approximately 7 cm. annual increments during the main 

 growmg period of that species, and in Notothenia ramsayi it seems fairly certain that increments of the 

 order of 10, 8 and 6 cm. accrue during the first three years of life, With perhaps a 4 cm. addition in the 

 fourth year. It is probable that in this species, as in hake, the males grow much more slowly than the 

 females once maturity is reached. Among species of intermediate size, 7 (? + ), 5 and 4 cm. increments 

 durmg the first three years of life are suggested for Notothenia jordani. There is a possibility that 

 further work might show Champsocephalus esox to be a fish of exceptionally rapid growth. 



