92 DISCOVERY REPORTS 



Atlantic deep water which enters the Indian Ocean between the Cape of Good Hope 

 and 56 S and partly of north Indian deep water. 



The salinity, temperature, and oxygen distribution in the Atlantic-Antarctic basin 

 as a whole, points to the existence of a large cyclonic circulation in the deep layer similar 

 to that which has already been described at the surface. The north Atlantic water which 

 flows eastwards into the Indian Ocean and is joined by north Indian water, turns first 

 southwards, then westwards along the edge of the Antarctic Continent and northwards 

 along the east coast of Graham Land, and finally back to the east along the northern 

 side of the basin. As it flows along this path the temperature and salinity of the current 

 decrease continuously owing to mixing with the colder and less saline surface and 

 bottom waters. 



Section 8 (Plates XIII-XV) crosses the cyclonic circulation near its eastern end, and 



the high salinity of the deep water close to the continent (3474-3475 °/ 00 in 59-64 S) 



shows that the deep water finds a much easier path towards the south in this region 



than it does farther west. At Sts. 850 and 851 there are, however, still some indications 



that the current is interrupted by eddies of colder water from the west. The temperature 



and salinity observations at St. 850 point to the existence of a particularly active eddy: 



the evidence which the section gives may perhaps be slightly exaggerated, since the 



observations at Sts. 155 1-2, which have been included in the section to bridge the large 



gap between Sts. 850 and 851, were made in a different year. It is also possible that 



the eddy is not a part of the major cyclonic circulation of the Atlantic- Antarctic basin, 



and it may be entirely or partly a local movement caused by the deep water flowing 



over the very uneven bottom in this region. The bottom topography is not well known, 



but it has been found to be very irregular, and the soundings change as much as 2000- 



3000 m. over very short distances. The distribution of temperature in the bottom layer 



(see pp. 112,1 13, and Plate XLIV) indicates that an S-shaped ridge connects the southern 



part of the mid-Atlantic ridge with the Marion Island-Crozet Islands ridge, and such a 



ridge might be the sole cause of the eddy. A second and weaker eddy was found at 



St. 851 in 56 22' S, 37 22' E, and the salinity, temperature and oxygen charts in Figs. 



19-22 suggest that it marks the eastern end of the cyclonic movement; farther east the 



warm deep water appears to flow southwards without interruption. 



The deep current towards the west along the continental slope south of the Atlantic 

 Ocean has at first a width of about 500 miles ; the isotherms and isohalines in Figs. 19-22 

 suggest that all the deep water south of 60-62 S in the region between 15 and 5 E, 

 flows towards the west. In 2-4 E the current divides on the Maud Bank, on which 

 Isachsen (1934, p. 161) found a sounding of only 1200 m. in 65 S, 2° 35' E: a longi- 

 tudinal section drawn by Mosby (1934, figs. 17-19) in 5 E, based on the Meteor data 

 (Wiist, 1928, pis. xxxiii, xxxiv) and the Norvegia data, shows that the warm water at 

 the Meteor St. 130 in 64 S is separated from the main body of the westward current 

 farther south by a region of colder water. Mosby (p. 38) suggests that the indication of 

 the splitting of the current was only a misrepresentation of the actual conditions, brought 

 about by combining data from different years, but he probably was not aware of the 



