188 DISCOVERY REPORTS 



During the period of this gale the wind records at Walvis Bay do not show any particularly stormy 

 trend, and so it appears, as indeed one might expect, that the gale was solely due to increased velocity 

 of the trade wind. 



In interpreting the results certain assumptions must be made. First, surface heating must be 

 assumed to have been regular within the region in question. Secondly, the turbulent mixing of the 

 upper water layers must be assumed not to have had any great effect in the lowering of temperature 

 in the surface waters. This latter seems justifiable since the depth of the mixed layer at WS 1079 

 before the gale was 50 m., and within this the temperature was nearly uniform. 



Fig. 36. Effect of wind and surface-temperature. The ship's track, stations and isotherms are shown in black for the west- 

 ward passage along the line of stations off Walvis Bay. While at the seaward end of this line, a S.S.E. gale was encountered 

 which blew for about 24 hr. at Beaufort force 8-9. Subsequently the ship returned eastward to Walvis Bay, completing 

 WS 1081 en route, and the track and isotherms on the return journey are shown in red. Probable surface water-movements 

 during the interval are indicated by the arrows. 



If these assumptions are correct we can imagine that the clockwise rotation of the isotherms offshore 

 was indeed the result of a clockwise water-movement. The change inshore suggests that an offshore 

 movement south of the line had been accompanied by a landward compensation flow in two wedges, 

 one moving south through the position of WS 1081, and the second right inshore at Walvis Bay. 



The mechanism of upwelling 

 With the accumulation of further theoretical knowledge of the movements of ocean currents it was 

 soon to be realized that upwelling was probably a far more complex process than a simple uniform 

 uplift of the subsurface layers. Without any factual data at his disposal, Bobzin (1922), in the third 

 part of his work, viewed the movements of the Benguela current from a theoretical angle. His 

 deductions, based on a consideration of the general principles of the movements of ocean currents, 

 suggested that the Benguela current existed in the form of a left-handed screw, proceeding towards 

 the equator. This conclusion, in view of later work, seems to have been almost a premonition. 



Gunther (1936) found from the 'William Scoresby' observations in the Peru current that the up- 

 welling was a localized phenomenon, and took place within a series of well-defined horizontal eddies, 

 while Defant (1936) working on the 'Meteor' data from the South-west African region developed 



