244 DISCOVERY REPORTS 



The Steep slope of the shelf edge is almost certainly one of the factors involved. These complications 

 in the coldest part of the Falkland current occur mainly to the eastward of the trawling area, and 

 therefore need not concern us further here. 



Along the mainland coast the speed of the Falkland current is greatly reduced and close inshore 

 southerly movements of surface water may predominate. Consequently the water close to the main- 

 land is warmer than that over the plain of the shelf. A definite counter-current close to the beach is 

 set up in the summertime, which may flow (more intermittently) at other seasons also. These inshore 

 conditions are clearly indicated by the direction of the isotherms in Klaehn's charts (191 1, pi. 35). 

 The warmer inshore water does not result from any southward translocation of subtropical water, 

 like the Brazil current offshore. The latter begins to swing away from the land well to the north of 

 our area, usually in about 30° S. The warmer inshore water is formed by 'warming-up' of sub- 

 Antarctic water, owing to the slower rate of advance on the left flank of the Falkland current. It may 

 be described as ' old shelf-water '. 



In the area of the trawling surveys there are only these two main hydrological features to be 

 visualized, the northward flow of the Falkland current over the plain of the shelf, colder and faster^ 

 offshore along the shelf edge ; and the old warmer water close inshore with a more or less definite 

 southerly trend. 



It is helpful in considering the distribution of fishes to gain some idea of the annual cycle of tem- 

 perature of the water. This enables some direct comparison to be made with conditions on better- 

 known fishing grounds elsewhere. Here we are handicapped by the fact that our three surveys were 

 made at different seasons in different years. Klaehn's very thorough averaging of results from many 

 years' observations, at a time when the region was much more important for traffic,^ provides the 

 general picture we need, though it deals with surface waters only. From a careful check of Klaehn's 

 charts against our own results it appears that the three years in which our own surveys took place 

 must have conformed closely to the 'average' conditions depicted by him. Comparing observed tem- 

 peratures obtained with the best modern apparatus on our surveys with Klaehn's monthly means, 

 we find a resultant mean error of -0-26° C. in autumn, -0-40° C. in winter and -0-07° C. in 

 summer. The range of error introduced by 'assuming' Klaehn's mean values instead of our own 

 would be -i-oi to +0-49° C, -1-35 to +o-8i° C, and -3-21 to +1-95° C. respectively. Some 

 error must be introduced by the time interval between our individual observations and the middle 

 of the month, the rest may be confidently attributed to diurnal variations, as indicated by the dis- 

 crepancy being greatest in summer. Klaehn worked on ten years' records from sailing ships (doubtless 

 mainly nitrate clippers) and four years' records of the German Admiralty. We know that two years 

 (1896-7 and 1906-7) out of the total studied by him were exceptionally cold, with icebergs drifting 

 far north in the Falkland current (Krummel, 191 1, p. 606, fig. 172). Such conditions are rare, and 

 this fact alone is probably sufficient to account for the slight tendency of Klaehn's values to be lower 

 than ours, especially in winter. More important still, a comparison between the observed differences 

 m surface temperature between successive pairs of our own observations, and the differences that 

 would be expected from Klaehn's results revealed a close correlation {r= +0-85, with P much less 

 than o-oi). It seems clear, therefore, that Klaehn's results give a very adequate picture of the ' average ' 

 conditions, which will give a more satisfactory idea of the annual cycle of surface temperature than 

 any scheme of plotting our less numerous and more scattered observations with interpolations for 

 the gaps. 



1 Some 13 sea miles per day according to Klaehn (191 1). 

 sequently""^"^ °^^" ^^"' °^ '^^ southern ocean, this region was far better known during 'the last days of sail' than sub- 



