148 DISCOVERY REPORTS 



never observed in the open ocean, where the decrease very often is small in the surface 

 layers and greater at some intermediate depth. Hence, a pure drift current must always 

 be present when a wind blows, and within this a transport of water takes place, which 

 is directed 90° cum sole from the direction of the wind. Fjeldstad (1930) has shown that 

 the total transport can be derived from the wind resultant and, therefore, all the irregu- 

 larities in the wind systems need not be considered. Because of this transport vertical 

 circulations must be developed, partly because of the limitations of the oceans and 

 partly because of convergences and divergences in the wind systems. 



The vertical circulations, which are maintained by the wind systems, would lead to 

 changes in the distribution of density, which, as previously shown, would further develop 

 the solenoid field which the wind builds up, and thus further increase the velocity of the 

 convection currents. Since the distribution of density in the oceans appears to have a 

 stationary character, it follows that the eff'ect of the vertical circulation, which is caused 

 by the wind, must be counteracted by other factors which influence the distribution of 

 density. The great importance of these other factors should thus be to maintain a certain 

 distribution of density and, therefore, when considering the total development of the 

 ocean currents, they must be given the same weight as the effect of the wind. These 

 factors also contribute to the development of vertical circulations, especially because 

 cold and heavy water sinks in high latitudes. 



It has been shown that vertical circulations probably exist because of the action of the 

 wind, and several systems of this nature are well known. The so-called "upwelling" 

 along the west coasts of Africa and America result from the action of the wind. The 

 convection currents along these coasts are directed such that heavy water is accumulated 

 along the coast but no vertical component would be present if the prevailing wind had 

 not a component along the coast, which caused a transport of the surface layers away 

 from the coast. 



Little attention has been paid to other systems of vertical circulation, but a very 

 interesting example is found within the current around the Antarctic Continent. This 

 current carries two typically different water masses, the Antarctic and the sub-Antarctic 

 water (Clowes, 1933). These two water masses are separated by a boundary surface 

 which at sea-level is recognized as the Antarctic Convergence or the Antarctic Ocean 

 Polar Front in the terminology of the German oceanographers. The boundary surface 

 between the two water masses takes in this case the place of the coast and makes possible 

 a vertical circulation within the Antarctic water. Before dealing with this circulation it 

 is necessary to consider the Antarctic Convergence more closely. 



Fig. I gives in the lower part a section of the density of the water along the meridian 

 30° W according to the Discovery observations. The boundary surface between the 

 Antarctic and the sub-Antarctic water masses is indicated by a dashed line, and this 

 boundary surface reaches sea-level in about latitude 50° S, where the Antarctic Con- 

 vergence is situated. It is seen that the light sub-Antarctic water covers the heavy 

 Antarctic water in the form of a wedge. In the upper part of the figure the inclinations 

 of the sea surface and the 1000 decibar surface, relative to the 3000 decibar surface, are 



