GENERAL ACCOUNT OF THE FISH FAUNA 251 



direction are not involved, observations at different seasons within any one of these regions are 

 obviously more fairly comparable than those over the whole area. In the specific distribution lists, 

 the regions to which individual stations belong are indicated by printing the numbers of northern 

 stations in heavy type, intermediate stations in ordinary type, and southern stations in italics. 



Secondly, the reversal of the seasons in the southern hemisphere is indicated by beginning all time 

 scales on i July, comparable to i January in the northern hemisphere. It should also be remembered 

 that when the looser seasonal expressions 'spring', 'summer', etc., are used, a similar reversal is 



'Twing to the shght gradient on the shelf, distance from the mainland coast is more significant in 

 relation to seasonal changes in distribution than changes in depth of water inhabited by migratory 

 species This figure has been calculated (in sea miles) for all the stations and is tabulated m 

 Appendix I Conditions round the Falkland Islands themselves are peculiar, and they do not appear to 

 be regarded as a coast at all by the main migratory species, which are almost absent f»-om their 

 immediate vicinity. I have therefore calculated the distance from the nearest point of the Falkland 

 Islands also, for those stations that fall within 100 sea miles of them. Beyond that distance migratory 

 fishes seem to proceed towards the mainland (in due season) as though the Falkland Islands did not 

 exist even though the mainland may be twice as far distant. The exact distances from the Falkland 

 Islands of all other stations has not, therefore, been calculated; they are tabulated merely as more 



than 100 miles'. • ,• ■ r>- 



Abbreviated descriptions of gear are the same as those used throughout the station lists m Discovery 

 Reports. The meanings of those used in this work have already been given m the section on field 

 methods. 



GENERAL ACCOUNT OF THE FISH FAUNA 



A list of all the species of fishes recorded on the Patagonian Continental Shelf and immediately 

 adjacent coasts is given in Table i. This also shows which species were obtained during the surveys 

 in ' Trawl + accessory nets ', in ' Other gear ' and by shore parties. Most of the specimens recorded in 

 the last category were obtained by Mr A. G. Bennett and Dr J. E. Hamihon. I have not included 

 some recent records of Pozzi and Bordale (1935) which are mentioned by Norman (1937, PP- HS-^)- 

 Some of these require confirmation. The list follows that of Norman (loc. cit., pp. 143-5) but omits 

 species recorded only from the west coast (columns A and B of Norman's list). The historical aspect 

 of the growth of our knowledge of the fish fauna is admirably dealt with by Norman (1937, PP- i37-4^). 

 and from his account the debt we owe to earlier expeditions-British, French, Swedish and American 



— can be assessed. . . r 



From Table i it can be seen that ninety-five species are recorded from the region, and specimens ot 

 seventy-eight of these were obtained by ships or shore parties in the course of the Discovery investi- 

 gations. Also an undoubted basking-shark was observed. The records of two of the species that we 

 did not capture seem somewhat doubtful: Bunocottus apus Kner is based on a single specimen said to 

 have come from the Burdwood Bank, no other Cottidae are known from the region and Kner s 

 description does not agree with his figure (cf. Norman, 1937. P- HS)- Alphestes afer (B och), a small 

 serranid common in the West Indies, has a normal range extending from Cuba to Brazil (Jordan and 

 Eigenmann, 1890, p. 35°). It is clearly a tropical species, so that its occurrence south of 42 ^ would 

 be most extraordinary, and I am unable to trace the authority upon which Norman placed it in the 

 Patagonian list. Among the other species not taken on the surveys, Notothenia tngramma Regan and 

 Crossostomm fasdat^^s (Lonnberg) are known only from their unique holotypes, and it is possible that 



