THE PACIFIC OCEAN 



105 



a salinity of about 34.66 per mille, and it shows a high 

 oxygen content since the water on the shelf is nearly 

 saturated with oxygen. Along the Antarctic coast of the 

 Weddell Sea the flow of the water is directed toward the 

 west and the westward motion of the waters can be 

 traced as far east as the region of Enderby Land. This 

 westward flow represents the southern part of a big eddy 

 which characterizes the entire Weddell Sea region. 



The observations from the Australian Antarctic ex- 

 peditions and from L. Christensen's expedition with the 

 Thorshavn show that sinking of water from the continen- 

 tal shelf does not contribute materially to formation of 

 bottom water within the entire region from Enderby Land 

 and eastward to Drake Passage and they show, further- 

 more, that the flow of the deep water is directed toward 

 the east within the entire region. From observations at 

 a few stations it is evident that water from the shelf in- 

 termittently sinks to great depths but in small quantities 

 only, for which reason the character of the bottom water 

 is only slightly influenced by these processes. The 

 previous hypothesis of the writer, that bottom water was 

 formed all around the Antarctic Continent and that a flow 

 of bottom water toward the west took place in every re- 

 gion, must therefore be abandoned. It is probable that 

 the surface waters near the continent flow toward the 

 west, but within the deep water there evidently exists an 

 Antarctic circumpolar current which flows toward the 

 east and follows the continent (except in the region of the 

 Weddell Sea) as far east as Enderby Land, where a big 

 eddy occurs on the southern side of the circumpolar cur- 

 rent. 



The characteristic properties of the water masses 

 within this circumpolar current are mainly determined 

 by the deep-water flow in the Atlantic Ocean, including 

 the Weddell Sea area. The deep-water flow within the 

 Atlantic Ocean has recently been discussed by Wiist 

 (1935) who has shown that three areas exist within which 

 the surface waters attain such a high density that they 

 must sink and contribute to the renewal of characteristic 

 water masses at great depths. One of these areas is 

 represented by the Mediterranean. Water of very high 

 salinity flowing out from the Mediterranean mixes with 

 Atlantic water and spreads toward the north and the 

 south, where it can be traced as an upper deep water. A 

 second area is found in the waters between Iceland, 

 Greenland, and Labrador. Within this area Atlantic 

 water of relatively high salinity is mixed with Arctic 

 water and cooled to such a low temperature that in some 

 localities water is formed which is of uniform density 

 from the upper layers to the bottom. Here water from 

 the upper layers may sink to great depth and contribute 

 to the renewal of the Atlantic lower deep water (mittleres 

 Tiefenwasser, according to Wust's terminology) which 

 can be traced to latitude 55° south. Within this region 

 or farther north, conditions may favor the development 

 of a water of lower temperature and lead to formation of 

 the bottom water of the North Atlantic, but this type of 

 water does not spread to any considerable distance and 

 is, therefore, of minor importance. The third area is 

 within the Weddell Sea, where the Antarctic bottom 

 water is being formed in the manner which has been de- 

 scribed. This bottom water spreads toward the north 

 and can be traced to latitude 40° north. 



Within the Atlantic Ocean we find, therefore, an 

 "active" deep-water circulation, especially between the 

 sea to the south of Greenland where the Atlantic lower 

 deep water is formed, and the area of the Weddell Sea 



where the Antarctic bottom water originates. No such 

 "active" deep water circulation is present in the other 

 oceans. In the Indian Ocean water from the Red Sea 

 spreads at moderate depths, but is of much less impor- 

 tance than the Mediterranean water in the Atlantic. 

 Water corresponding to the Atlantic lower deep water is 

 not formed in the Indian Ocean nor is Antarctic bottom 

 water formed south of the Indian Ocean. Within the en- 

 tire area of the Pacific Ocean no renewal of any type of 

 deep water takes place. 



The water masses of the Antarctic circumpolar cur- 

 rent are, as already mentioned, formed by mixing of 

 Atlantic deep water and Antarctic bottom water. Wiist 

 has shown that such processes of mixing take place to a 

 great extent in the Atlantic Ocean, and he has computed 

 the percentage amount of true Atlantic deep water or 

 true Antarctic bottom water in the layers of the Atlantic 

 Ocean. The two types of water are still characteristi- 

 cally different within the circumpolar current in the 

 southern part of the Atlantic Ocean, but when carried 

 toward the east by this current the differences disappear, 

 owing to processes of mixing, and to the south of Aus- 

 tralia we find water of a very homogeneous character 

 which can be described as a special type of water, the 

 Antarctic circumpolar water. The temperature of this 

 water lies between 0° and 2° and the salinity between 

 34.68 and 34.74 per mille. 



This water flows, as already stated, around the en- 

 tire Antarctic Continent and follows the continental slope 

 except in the region of the Weddell Sea where there is a 

 large eddy south of the circumpolar current. This is 

 evident from the observations of the Australian Antarc- 

 tic expedition, and Clowes (1933) has convincingly shown 

 that the flow through Drake Passage is directed from the 

 Pacific to the Atlantic Ocean. Accurate determinations 

 of the oxygen content within the circumpolar current 

 might confirm this conclusion. From the Meteor obser- 

 vations (in 1926) Wiist finds in the Weddell Sea region an 

 oxygen content of 4.6 ml/L at a temperature of l.°6, and 

 of 5.6 ml/L at a temperature of -0°6. The o.xygen ob- 

 servations on the Australian Antarctic expedition and 

 L. Christensen's expedition with Thorshavn, and obser- 

 vations from the Drake Passage on board Discovery II 

 in 1931 indicate a decrease of the oxygen content of the 

 deep water from the region north of the Weddell Sea and 

 eastward to Drake Passage. Within the Antarctic cir- 

 cumpolar water, the oxygen content increases toward the 

 bottom and the temperature decreases. Thus, a relation 

 exists between the oxygen content and the temperature 

 and, on an average, the oxygen content is nearly a linear 

 function of the temperature. 



In the Weddell Sea region (1926) 



02 = [4.42 + 0.45(2°-t)]ml/L 

 In the Indian Antarctic Ocean (1929-1930) 



02 = [4.18 + 0.50(2°-t)]ml/L 

 In the Drake Passage (1931) 



02 = [3.95 + 0.45(2°-t)]ml/L 



The Meteor observations in the Drake Passage in 1926, 

 however, show very nearly the same oxygen content as 

 the water of similar temperature and salinity to the 

 north of the Weddell Sea. 

 Drake Passage (1926) O2 = [4.35 + 0.45(2°-t)]ml/L 



Thus, the evidence is conflicting and at present it can 

 only be stated that a majority of observations indicate a 

 decrease of the oxygen content of the deep water in an 



