ii 4 DISCOVERY REPORTS 



temperature of the bottom water has increased still further and its oxygen content has 

 decreased. The salinity, however, shows no increase, and since the higher temperature 

 suggests that the bottom current has been mixed with warm deep water the low salinity 

 is difficult to explain. The current may have been diluted with a mixture of shelf water 

 and warm deep water but the low oxygen content of the bottom water argues that 

 there is no such addition and the data as a whole show that the volume of surface water 

 entering the current could only be very small. The salinity difference involved is only 

 about o-oi °/ 00 and not more than 0-02 °/ 00 , and although it is unlikely that such an 

 error would enter systematically into the titrations, the difference needs some con- 

 firmation before it can be taken seriously. 



In section 13 (Plates XXVII I-XXX), the only deep observations were made in the 

 Tasman Sea and in a deep basin near Macquarie Island. The high temperature and low 

 oxygen content of the bottom water in these basins suggests that the bottom current 

 only finds its way into them after being substantially mixed with the warm deep water. 

 It is again remarkable, however, that the relatively large increase of temperature and 

 decrease of oxygen content has been achieved with only a small increase of salinity, 

 and it seems that the bottom water must have been mixed with deep water whose 

 salinity has been reduced almost to that of the bottom current by turbulent mixing; 

 an examination of the sections and a knowledge of the exceptionally rugged nature of 

 the bottom topography in the region suggest that this is not impossible. A rough 

 calculation shows that deep water such as that found between 400 and 800 m. at 

 St. 912 and between 1500 and 2000 m. at St. 919 in section 13 when mixed with the 

 bottom water at St. 905 in section 12 would give rise to water with the same properties 

 as the bottom water at St. 919. 



Our observations in the sub-Antarctic and subtropical parts of the Indian ocean serve 

 on the whole to confirm the conclusions reached in a recent discussion of the bottom 

 currents of the Indian Ocean by Wiist (1934). This work suggested that the Indian 

 Ocean was divided by a ridge into eastern and western basins similar to those of the 

 Atlantic Ocean. Wiist draws the ridge northwards from the Kerguelen plateau through 

 New Amsterdam and Rodriguez Islands to the Chagos and Maldive Islands, but more 

 recent soundings by the Murray expedition (not yet published) indicate that a sharper 

 connection exists through the Carlsberg ridge (Schmidt, 1932) to Socotra. The bottom 

 water enters the western basin through the Kerguelen Channel, between the Crozet 

 Islands and Kerguelen, and it enters the eastern basin through a gap in the Indian- 

 Pacific cross-ridge that is traversed by section io 1 ; the depth of the passage seems to be 

 just over 4000 m. 



The observations at the southern end of section 14 (Plates XXXI-XXXIII) show 



that the bottom water just north of the Ross Sea also has properties which indicate 



that it belongs to the eastward current from the Weddell Sea, and they give a clear 



indication that there is no stream of cold poorly saline bottom water sinking from the 



1 See List of Oceanic Depths received at the Admiralty during 1932 (Hydrographic Department, 1933), 

 and Deacon (1934, p. 131). 



