8 DISCOVERY REPORTS 



A comparison of these data with the observations made farther south — for the most 

 part south of 6o° S— in the Pacific Ocean, indicates that the surface water in the 

 northern part of the zone is the first to be raised above its winter temperature. The 

 earlier formation of the stratum in the north is no doubt caused by the greater amount 

 of radiation falling on the northern part of the zone. 



It is also natural to conclude that the water in the southern part of the zone will be 

 the first to cool to its winter temperature at the end of summer. The first sign of the 

 approach of winter was noted in the middle of March at St. 1 1 54 (section 7, Plates X-X 1 1 ) 

 in 69 20' S, 9 34' E. The surface water was cooled to a temperature of -1-57° C, 

 whilst the water at 60 m. — probably the remains of the warm surface stratum of the 

 previous summer— was not colder than -i-io° C. The observations made at this time 

 north of the Antarctic Circle suggest that the surface water had not begun to cool, and 

 even as late as the middle of March the radiation and conduction of heat from the 

 surface appears to be balanced by the heat which is being absorbed. At Sts. 1 156 and 

 1 1 58 the surface water was 2-19 and 2-39° C. warmer than the cold stratum, and the 

 low salinity of the surface water at St. 11 58 suggests the nearness of melting ice. 



The observations in sections 8 and 9 (Plates XIII-XVIII) in the Indian Ocean show 

 that by the second half of April and the first half of May all the Antarctic water was 

 cooling. As the surface water loses heat — principally by radiation— it becomes heavier 

 and sinks to mix with the deeper water. At first the mixing is confined to the surface 

 stratum, but as the density of this water approaches that of the cold stratum, it ex- 

 tends throughout the whole layer which in time becomes homogeneous. At the stations 

 south of 63 ° S near the end of April the water was coldest at the surface, but at those 

 farther north the surface stratum was still i-i to 1-7° C. warmer than the cold stratum. 

 South of 59-60 S in the region south of Australia and New Zealand the surface layer 

 was cooled right through to an even temperature ; but farther north, although the ob- 

 servations were made in the second half of May and June, the surface water was still 

 warmer than the lower stratum. The difference was most marked near the convergence : 

 at Sts. 883 and 891 (sections 10 and 1 1 , Plates XIX-XXIV) the first 100 m. of water was 

 1 -6 to 1 -3° C. warmer than the lower stratum. Such a contrast between the two strata so 

 late in the year is probably due partly to the existence of different currents in the two 

 strata; either there is a southward movement at the surface, or the northward move- 

 ment is strongest in the cold stratum. 



Large temperature differences between the two strata have been noted most frequently 

 in the region between the Falkland Islands and South Georgia. They appear to be caused 

 by a movement of sub-Antarctic water towards the south or east ; the movement does 

 not terminate abruptly where the bulk of the Antarctic water sinks but carries a surface 

 stratum of sub-Antarctic water 100-200 miles 1 farther south. At St. 1027 in section 3 

 (Plate IV) the average temperature of the first 50 m. of water was 3-53° C. as early as 

 the middle of November, but that of the water between 100 and 200 m. was only o-68° C. 

 Similar differences were found at Sts. 634, 635, and WS 519 (Station List, 1932) in the 



1 Throughout the report distances are expressed in nautical miles. 



