VERTICAL CIRCULATION IN THE OCEAN 143 



and that in the deep water the isosteres are nearly horizontal. This implies that the 

 greatest number of solenoids in the sense of Bjerknes are present in the upper layers and 

 that very few or no solenoids are found in the deep water. Consequently the velocity of 

 the convection currents decreases rapidly with depth and approaches zero at great 

 depth, contrary to the velocity of the slope current, which remains constant from the 

 surface to the top of the lower layer of frictional resistance. 



Ekman, furthermore, points out that in a sea in which the density increases with 

 depth, a pure drift current must give rise to a convection current, running generally in 

 the direction of the wind. In order to illustrate this effect of the wind we may again 

 consider the channel around the earth, but now we will assume that the water is not 

 homogeneous but that the density increases with depth. The immediate effect of a pure 

 drift current will be to transport the light surface water towards the left-hand side of 

 the channel and, therefore, near the surface the isosteres cannot remain horizontal but 

 must soon slope downwards from south to north. In the upper layers the light water is 

 accumulated on the left-hand side of the wind and we get a solenoid field which must 

 cause a current in the direction of the wind, but the velocities within this current de- 

 crease downwards since the inclination of the isosteres decreases. At the same time we 

 may get a piling up of the water at the northern boundary, and as a result of this a slope 

 current may arise, which also runs in the direction of the wind but remains constant 

 down to the top of the lower layer of frictional resistance. 



Similarly we may consider a circular wind system within an area which is situated 

 completely in one hemisphere. If the circulation is contra solem the surface layers are 

 driven away from the centre and a field of solenoids is built up which results in a current 

 in the direction of the wind. If the circulation is cum sole the light surface water will be 

 carried towards the centre and a density distribution is brought about which gives rise 

 to a convection current circulating cum sole. 



These considerations show the character of the currents which may be produced in 

 homogeneous and non-homogeneous water under the action of the wind, and also that 

 conditions become much more complicated when the water is non-homogeneous. In 

 order to undertake a complete analysis of the ocean currents one has, furthermore, to 

 consider changes in the distribution of density, which are caused by such factors as 

 heating and cooling, evaporation, etc. The problem becomes so complicated that at 

 present it cannot be made the subject of any mathematical analysis, but it seems possible 

 to follow some simple lines of reasoning which lead a step further and which give some 

 indications of the structure of the circulations in the sea. 



In the first place it is of much interest to ascertain whether any evidence is present 

 for the existence of slope currents in a non-homogeneous sea. It is not a priori given 

 that such currents are developed. As will be shown below, under the action of the wind 

 on the surface of a channel of uniform depth running around the world, a stationary 

 current can exist only in the absence of a slope current, and generally it is possible that 

 the principal effect of the wind is to build up a solenoid field which gives rise to a con- 

 vection current and not to bring about a piling up of the water along a boundary. 



