ioS DISCOVERY REPORTS 



It seems very doubtful, however, that such sinking of surface water ever takes place ; 

 the region described has been shown to be a divergence region along the axis of an 

 extended cyclonic circulation (see pp. 92-94) rather than a convergence region as 

 supposed by Wiist, and the bottom water is more likely to well-up towards the surface 

 than the surface water to sink. The upwelling of the bottom water would explain 

 the great vertical extent of the cold water in the region. It is also reasonable to argue 

 from the temperature, salinity and oxygen distribution in the bottom layer that the 

 warmer types of bottom water are formed by the mixing of the bottom current which 

 flows from the south-west corner of the Weddell Sea with more warm deep water. 



Although the large area described by Wiist seems most certainly to have no re- 

 semblance to those described by Nansen as sources of Arctic bottom water, there is 

 some evidence to show that there is a small region between the South Shetland Islands 

 and the South Orkney Islands for which the analogy is more just. In this region there 

 are only traces of warm water in the deep layer. The observations at St. 637 in 6o° 00' S, 

 49 28' W and St. 638 in 6i° 00' S, 49 48' W (Station List, 1932) indicate that the 

 temperature of the deep water does not rise above 0-05° and —0-17 C. At St. 639 in 

 61 ° 58' S, 51 59' W the layer is slightly warmer, having a temperature of 0-34° C. 

 Another striking feature of the region, shown by these observations, is the very uniform 

 and low salinity of the deep and bottom waters : below 600 m. the salinity of the water in 

 the whole region covered by the three stations only varies between 34-57 and 34-64 % . 

 Both the temperature and salinity distribution suggest that in winter the whole water 

 column from surface to bottom may be homogeneous. This seems to be especially true 

 of the water at St. 637, where although the observations were made in March and the 

 surface water had a temperature of 0-52° C. — at least 2 C. above its probable winter 

 temperature — the salinity of the surface water was as much as 34-35 °/ 00 — only 0-28 °/ 00 

 less than the maximum at a depth of 2500 m. 



The conclusion that the region is one in which there is active convection between 

 the surface and bottom layers in winter, so that the surface water is carried directly 

 into the bottom layer, is definitely supported by the high oxygen content of the deep and 

 bottom waters ; although the observations were made in March, and the deep water, if 

 it sank from the surface during the previous winter, had spent several months shut 

 off from a source of oxygen, its oxygen content at Sts. 637 and 638 had not fallen below 

 5-19 and 5-38 cc. per litre. At St. 639, where there was warmer water in the deep layer, 

 it had, however, fallen to 4-36 cc. per litre. 



The low salinity of the bottom water — less than 34-64 °/ 00 — shows that it was not part 

 of a bottom current from the Weddell Sea, where the salinity of the bottom water is not 

 less than 34-66 °/ 00 . It seems, therefore, safe to conclude that it was formed in situ 

 during the winter. The bathymetric chart given by Herdman (1932, pi. xlv) shows that 

 it is not confined to a deep basin shut off from the deep water of the neighbouring seas 

 by shallow submarine ridges; St. 637 lies north of the Scotia Arc and St. 638 and 639 

 south of it, but the bottom waters appear to be in free communication with those of the 

 Scotia Sea and Weddell Sea. The small difference between the bottom waters on either 



