OBSERVED DISTRIBUTION OF TEMPERATURE AND SALINITY 169 



to local surface-heating, but on the other hand the bathythermograph observations definitely suggest, 

 although not conclusively, that this feature is part of an extensive detached eddy between ' oceanic ' 

 and 'coastal ' types of water. If this interpretation is correct the cooler water outside this warm area 

 would combine to give the eddy an overall diameter of some 120 miles, and extending in depth to 50 m. 



At Cape Frio, in 18 30' S., a temperature of 15-1° C. was recorded about 7 miles offshore, but 

 north of this yet another patch of cooler water was present, extending up to the Cunene river mouth. 

 From there northwards, however, the temperature of the surface-waters increased steadily, until in 

 14 S. the tropical boundary was encountered and the temperature increased rapidly to values over 

 22° C. 



The distribution of temperature and salinity at a depth of 100 m. (Fig. 20) has the same general 

 characteristics as that at the surface, the cooler and less saline waters lying uniformly along the coast, 

 with the warmer more saline waters offshore. Although the smaller number of observations at this 

 depth are probably responsible for the smoothing out of the isotherms and isohalines, they still show 

 up quite clearly the pronounced eddy in the vicinity of the Orange river. The regions of lowest 

 temperature and salinity as shown by the io-o° C. and 34-80 % isolines are present to the south of the 

 Orange river and off Luderitz Bay. 



At 150 m. an entirely different pattern appears. This is not illustrated but along with the distribu- 

 tion at 200 m. it represents a transition to the distribution at 300 m. (Fig. 21). At 300 m., in spite of 

 the paucity of observations, it is possible to recognize that the distribution is roughly similar to that at 

 a depth of 200 m. on survey I. Here we find an apparently anomalous situation with the warmer and 

 more saline water nearest to the coast while the cooler water lies further offshore. There can be no 

 doubt that further to the west warmer water must have been present, and so it seems that on this 

 survey the cooler water at this depth must have been still more extensive and further displaced from 

 the shore than on survey I. This was concurrent with an increase in the current of warmer water 

 which runs south along the coast. 



Vertical distribution of temperature and salinity 

 An outstanding feature of the second survey in comparison with the first is the absence of the very 

 strong thermoclines which were found so frequently in March. At most stations, particularly those in 

 the offshore waters, there is a convection layer on the surface, and below this the temperature and 

 salinity decrease gradually into the deeper water. In several places where a well-defined convection 

 layer had apparently been present, surface heating has reduced the density and produced a thin stable 

 surface layer. The depth distribution of the convection layer on survey II is shown in Table 5. 



The Orange river line (28 30' S.). The low surface salinities inshore are a result of the fresh-water 

 inflow of the river, and they have produced a very stable surface layer at station WS 1050. Further 

 offshore, however, at WS 1051 a thin convection layer is present, and this deepens and becomes more 

 pronounced farther to the west. The vertical sections of temperature and salinity (Figs. 22 and 23) 

 contrast vividly with the conditions in March. The marked upward slope of the isotherms and iso- 

 halines towards the coast, combined with the active offshore transport of surface-water as suggested 

 by the salinity section, shows most clearly the characteristics of active upwelling. It will also be seen 

 that the isotherms and isohalines, while approaching the surface inshore, also rise towards the surface 

 in the proximity of the edge of the continental shelf. There seems to be clear indication, therefore, of 

 an additional divergent movement in this position. As will be shown later, this is probably associated 

 with the mechanism of the process of the upwelling. 



North of the Orange river line the tongue of warm offshore water at WS 1057 and 1058 has an 

 almost constant temperature in the upper 100 m. A relatively sharp boundary was present on the 



