ii2 DISCOVERY REPORTS 



passage through the mid-Atlantic ridge near the equator — and flows northwards as far 

 as 35 N and southwards to the Walfisch ridge in 20-35 S. 



Towards the east from the Weddell Sea the bottom water spreads freely across the 

 northern side of the Atlantic-Antarctic basin. Between the Weddell Sea and 30 E 

 the tendency of the cold water to keep to the northern side of the basin is most notice- 

 able, and there is a stream of warmer water along the edge of the Antarctic Continent. 

 This temperature distribution is a plain indication that there is a cyclonic circulation of 

 water in the bottom layer similar to those which have already been described in the 

 surface and deep layers. Like the deep water circulation, it is probably more complete 

 than that of the surface water, since the isotherms in the eastern part of the basin give 

 a more reliable indication of a southward movement to complete the cycle. 



The actual temperature and the other properties of the bottom water in the basin are 

 shown in sections 5, 6, and 7 (Plates VII-XII). The lower temperatures, salinities and 

 higher oxygen contents are found on the northern side, and the warmer, more saline, 

 and poorly oxygenated water on the south. The sections were not all made in the same 

 year, and they reveal the fact that relatively large variations may occur in the properties 

 of the bottom water. Although section 6 crosses the northern part of the basin some 

 distance to the east, and therefore farther away from the source of the bottom water than 

 section 5, the temperature of the bottom water was found to be slightly less and the 

 oxygen content considerably greater than those of section 5. It seems most probable 

 that the difference arises because the bottom water in the eastward current had dif- . 

 ferent properties, at the time when section 6 was made, from those which it had fourteen 

 months earlier at the time of section 5. Since the formation of the Antarctic bottom 

 water is probably much greater in winter than in summer, it can only be supposed that 

 the properties of the water in the eastward and northward currents from the Weddell Sea 

 will have a periodical change, the water flowing away from the sea in winter being colder 

 and containing more oxygen than that which leaves the sea in summer. It is hoped that 

 such differences will give a method of measuring the rate of flow of the bottom water, 

 similar to the method already used for the Antarctic intermediate current (Deacon, 

 1933, pp. 223-6), but the necessary close examination of all the data available has not 

 yet been made. 



There is not such a strong current of Antarctic bottom water towards the north in 

 the eastern half of the Atlantic Ocean as there is in the west, the difference between the 

 two streams being most probably the result of the existence of a deeper channel in the 

 west. Wiist in his temperature chart of the Atlantic Ocean and his vertical section along 

 the eastern Atlantic basin (1933, pis. i, vi) has assumed that there is a passage from the 

 Atlantic-Antarctic basin into the Agulhas basin, south of Africa, with a depth of about 

 4500 m., but our new data indicate that the passage may be shallower. In section 8 (Plates 

 XIII-XV) St. 849 was made in the Agulhas basin, whilst St. 850 was in the Antarctic 

 basin. The section shows that between the two deeps there is a steep ridge which acts as 

 a partial barrier to the northward flow of the bottom water. The ridge between the two 

 basins is probably not so well defined and so shallow throughout its entire length as it is 



