HYDROLOGY OF THE SOUTH ATLANTIC 187 



These movements are known to exist and the wind is probably almost entirely re- 

 sponsible for them. At the northern end of Graham Land a little water flows westwards 

 against the wind, into the Bransfield Strait. This may be due to the eff'ect of the earth's 

 rotation on the Weddell Sea current flowing towards the north-east out of the Weddell 

 Sea, or to the presence of a counter-current between the Weddell Sea current and the 

 Bellingshausen Sea current which also flows towards the north-east outside the Weddell 

 Sea water. The effect is a very minor one compared with the easterly movement in the 

 same region. 



There must, however, also be a northerly movement in the cold stratum of the 

 Antarctic surface layer, almost as strong as the northerly movement in the surface 

 stratum. This is shown by the amount of water sinking at the convergence which is 

 colder than the water in the surface stratum. Also, if the water found in the cold stratum 

 near the northern boundary of the zone were not continually renewed from the south, 

 the minimum temperature in the cold stratum would disappear, as a result of mixing 

 with the warmer water above and below it. Such a northerly movement can be explained 

 as part of the thermohaline circulation, and it is to this circulation, in all probability, 

 that north and south movements are mainly due. 



Because Antarctic surface water sinks below the surface at the Antarctic convergence 

 there must be a greater movement towards the north in the Antarctic surface layer than 

 there is in the surface layer in the sub-Antarctic Zone. This is not the result of a dif- 

 ference in the strength of the pure drift currents in the two waters, since, as is shown by 

 Table XII (p. 236), the wind is strongest north of the convergence. The convergence 

 is, moreover, sharp and the Antarctic water sinks suddenly, and there is no corresponding 

 sudden change in wind velocity which could produce a sharp convergence. These are 

 facts which afford additional reasons for believing that the northerly movement is not 

 purely a wind current but a convection current caused by density differences. 



The greater speed with which the surface water moves in summer cannot be explained 

 as the result of stronger winds in summer than in winter, since the winds are approxi- 

 mately of the same strength in both seasons (Schott, 1926, p. 224, and Ardley, 

 p. 235, infra). The increased speed can be explained by the thermohaline circulation, 

 in so far that there is a greater production in summer of surface water which is heavier 

 than that farther north ; but it is perhaps best explained as the result of a third factor 

 — the liberation of fresh water in summer. According to Schott the precipitation 

 between 50 and 60° S in the West Atlantic exceeds 1000 mm. per annum, and ac- 

 cording to Cherubim (1931, pp. 325-35) the amount of evaporation is just less than 

 300 mm. per annum. 



Ekman (1926, p. 261) has studied theoretically the production of ocean currents in 

 the sea which would be caused directly by the difference in precipitation in different 

 regions in the ocean. He has shown that the greatest velocity of such currents is 

 1-2 cm./sec, and that they will probably be only a fraction of this. In the Southern 

 Ocean, however, the snow which falls far south does not melt until summer approaches, 

 and then it is released quickly. There are the additional effects of drainage from the 



