i8o DISCOVERY REPORTS 



rapid in summer than in winter. Tlie principal evidence in support of this conclusion 

 is obtained from topographical maps, which show the height of the surface, calculated 

 from dynamic data, above some deeper isobaric surface which can be assumed to be 

 level. These maps indicate that the Antarctic surface currents in the Falkland Sector 

 are about twice as fast in summer as in winter. Further evidence, which is, however, 

 not conclusive, can be obtained from the movements of pack-ice. In the year 1930, 

 for example, there was no pack-ice north of South Georgia in August, September, or 

 October, and its appearance in November can be attributed to the speeding up of the 

 surface currents with the approach of summer. It would also be expected that the 

 Antarctic convergence would move northwards as the speed of the surface currents 

 increases; but the evidence we possess does not support this conclusion. Few obser- 

 vations have, however, been made in winter and those available are not sufficient to 

 disprove a seasonal movement of the convergence. 



Although it is only in the Falkland Sector that there are sufficient data to examine 

 the seasonal changes in the speed of the surface currents and in the extent of the layer 

 it can safely be assumed that there is a greater production of Antarctic surface water 

 in summer than in winter. This increased production can be explained by the great 

 additions of fresh water which the layer receives in summer from melting ice and as 

 drainage from the land. In winter these sources are for the greater part stopped, and 

 the activity in the layer diminishes. The origin of Antarctic surface water does not lie 

 entirely in the addition of fresh water to the layer in summer, or to a lesser extent in 

 winter. There must be further additions to the layer from a different source to maintain 

 its salinity. It will be shown later that besides the transport of Antarctic water towards 

 the north in the surface there is also a flow of Antarctic water northwards in the bottom 

 layer. To compensate for the loss of water towards the north by these two currents, 

 which are known to exist all round the Antarctic Continent, there must be a 

 return current towards the south in the intermediate depths. Also, between the cold 

 Antarctic surface layer and the cold Antarctic bottom layer there is a maximum tem- 

 perature in the warm deep layer. This maximum temperature, in water which lies 

 between two bodies of colder water, can be maintained only by a movement of water 

 southwards in the warm deep layer. A second source of Antarctic surface water is 

 therefore to be expected in the warm deep water. 



A certain amount of mixing must always take place between the two layers across the 

 discontinuity layer which separates them, especially in winter. It is principally to be 

 looked for in places where the density difference between the two layers is least, and in 

 places where for hydrodynamical reasons, based on the direction of winds, currents, or 

 the shape of land masses, the warm deep water is made to rise towards the surface. 

 Warm deep water has never itself been found at the surface, although it has been found 

 with its maximum temperature at a depth of only 100 m.: it is always covered with 

 Antarctic surface water. In section I the salinity of the water at the station made at the 

 edge of the pack-ice was greater than at those farther north ; its temperature was also 

 higher, and as will be shown in Fig. 18 (p. 204) its oxygen content was less. Each of these 



