96 DISCOVERY REPORTS 



and the southward movement of the North Atlantic deep water seems to end in about 

 6o° S east of the South Sandwich Islands and in 56 S south of Cape Town. 



The salinity, temperature, and oxygen distribution in the subtropical region of the 

 South Atlantic Ocean and south-east of the Cape of Good Hope in the Indian Ocean 

 shows that the North Atlantic deep water spreads towards the east as well as to the 

 south, but so far it has not been found possible to compare the strengths of the two 

 movements. The southward movement is probably caused by the demand for a current 

 to compensate the northward movements in the intermediate and bottom layers, and 

 the greater strength of these northward movements in the western half of the ocean is 

 reasonably supposed by Wust (1928, p. 531) to account for the greater strength of the 

 North Atlantic deep current in the western Atlantic basin (see Fig. 19, and Wust, 1926, 

 pp. 237-42). The effect of the earth's rotation alone would cause the current to be most 

 strongly developed in the eastern half of the ocean. 



The bending of the isotherms and isohalines towards the north in the region ad- 

 joining 30 W (see Figs. 19, 20) suggests that the southward movement of deep water is 

 weaker than it is farther west. The region is situated at the northern end of the deep 

 channel from the Atlantic-Antarctic basin to the Argentine basin (see Plate XLIV), 

 and the smaller southward movement is probably due to the influence of the stronger 

 northward bottom current. 



THE DEEP CURRENTS IN THE SOUTHERN PART OF 

 THE INDIAN OCEAN 



A brief description of the north Indian deep current, which Schott (1926), Moller 

 (1929), and Thomsen (1933) have shown to be formed by the sinking of highly saline 

 surface water in the Arabian Sea and adjoining gulfs, has already been given (pp. 90, 91). 

 The properties of the current, high salinity and temperature, and low oxygen content are 

 well known, but the distance to which it penetrates towards the south is still rather un- 

 certain. Moller (1933) regards the existing data as sufficient to show that it spreads 

 southwards as far as the Antarctic regions, even though Thomsen (1933) had shown 

 that the Dana observations along a line from the northern end of Madagascar to Ceylon 

 suggested that such a current might not exist south of the equator. A preliminary 

 examination of the data collected by the ' Discovery II ' along a line from Marion Island 

 to Socotra through the Mozambique Channel (Clowes and Deacon, 1935) shows that 

 although the most saline water in the deep layer south of 20 S belongs partly to an 

 eastward current from the Atlantic Ocean, the less saline water in the upper part of the 

 layer is probably derived largely from the north Indian deep current. In 20 S the 

 southward movement of the current appears to be partly obstructed by the northward 

 movement of Antarctic intermediate water, which lies at a much greater depth south of 

 20 S than in the northern part of the ocean ; but the oxygen section in particular sug- 

 gests that the current continues towards the south below the intermediate water and 

 above the Atlantic water. A comparison of Moller's longitudinal sections (1929, figs. 4, 

 5, 8, 9), based on the observations of the 'Valdivia' and 'Ormonde', with the section 



