VERTICAL CIRCULATION IN THE OCEAN 145 



the result, that the sea-level is higher along the Pacific coasts than along the Atlantic, 

 which can be brought into agreement with the distribution of density, the introduction 

 of the results of the levelling has raised new questions instead of solving old ones. 



There exist, on the other hand, observations which indicate that in the oceans the 

 piling up effect of the wind is negligible. The observations of the International Ice 

 Patrol around the Grand Banks of Newfoundland have been used for constructing 

 charts, showing the topography of the surface relative to some level at which the velocity 

 of the current presumably must be small. From these topographic charts the currents 

 at the surface can be derived, and since the tracks of icebergs have shown a remarkable 

 agreement with the computed currents, the latter must closely represent the actual flow 

 of the water. It is unnecessary to assume the existence of a slope current, and thus no 

 indication of a piling up of the water exists. 



A still more striking example is found in the Strait of Florida where, according to 

 Wiist (1924), the currents which are computed from the distribution of density agree 

 perfectly with the measurements by Pillsbury. There the sea-level rises 30 cm. over a 

 distance of 40 km. in a direction at right angles to the coast. This very big rise over a 

 short distance clearly shows the importance of the distribution of density to the in- 

 clination of the sea-level. We must, however, leave open the question as to the possible 

 piling up effect of the wind and the existence of slope currents, and we shall proceed 

 to a discussion of stationary wind currents in non-homogeneous water. 



Let us again consider the case of a channel running all around the earth in the southern 

 hemisphere. As a primary effect of the wind the light surface water will, as previously 

 stated, be transported to one side of the channel, a field of solenoids will be produced 

 and a convection current in the direction of the wind will commence. This convection 

 current must gradually become stronger as long as the pure drift current continues to 

 transport the light surface water across the channel. Stationary conditions can be 

 reached only when all transversal motion stops, because any transversal motion must 

 cause changes in the distribution of the density, if no other factors tend to maintain a 

 certain distribution. 



It is easy to find the conditions which must be fulfilled if stationary conditions are to 

 exist. The equations of motion have in the stationary case the form : 



9/> . , d I dvx 



dp d I dv^ 



•(I) 



where a represents the specific volume, p the pressure, v the coefficient of eddy viscosity 

 and V the velocity. The positive s^-axis is directed downwards. Placing the positive 

 j'-axis in the direction of the channel and the x-axis at right angles to the channel we 

 have: 



«!=/(-). -«|-- w 



