UPWELLING 203 



hin oder parallel zu ihm, aber nicht vom Land weg gehen und damit die Voraussetzung 

 fiir das Aufquellen von Tiefenwasser wegfallen " (p. 168). In support of his statement, 

 Schott brings forward the similarity between the temperature records of the two ships 

 ' Emden ' and ' Nitocris ' : these agree in showing that Arica lies in a region where the 

 inshore temperature approximates closely to that 100 miles offshore. 



In July 1 93 1, conditions were markedly different ; offshore temperatures being higher 

 and inshore temperatures lower than those given by Schott. While it is true therefore 

 that high temperatures were very close inshore, upwelling was also in active operation. 

 The ship's drift in a direction more or less parallel to the coast gave no hint of con- 

 vergence within a few miles of the shore in the Bight of Arica, the region where 

 convergence might be most expected (see Fig. 11). The paradoxical conclusion is 

 irresistible that some divergence from the coast was taking place in a region of 

 convergence. 



The differences between Schott's data and our own might be explained by the 

 variable nature of the current or by differences in the type of observations made. The 

 ' Emden ' and the ' Nitocris ' steamed along the current at an unspecified distance from 

 the shore, whereas the ' William Scoresby ' steamed across the direction of the current 

 and observations were possible at various distances from the coast. The curves plotted 

 by Schott resemble most the curve plotted by us in Fig. 34 for 2-5 miles from the shore, 

 and as the cool water at Arica occupied a very narrow band, it is possible that Schott's 

 data were collected outside the zone of upwelling. 



Murphy (1925), who knows the coast well, contributes some remarks upon the coast- 

 line. On p. 175 he states: "The coast of Peru. . . trending sharply to westward from 

 near the Chilean border, extends far into the ideal course of the Humboldt Current, and 

 forces the latter to become an actively impinging stream until it has passed the end of 

 the continental buffer at Point Parina." The data collected on the present survey bears 

 out Schott's statement that the west coast constitutes a single-sided divergence line; 

 the Peru current cannot therefore constitute an actively impinging stream. It will be 

 shown later (pp. 208-13) that impingement, or convergence, leads to a reduction or 

 cessation of upwelling. 



Trend of coast-line would probably rank as an important factor governing the 

 hydrological conditions if the winds off the Chilean and the Peruvian coast were 

 similar. But off Chile winds are predominantly southerly and off Peru south-easterly 

 (Fig. 4). Off both coasts they are therefore parallel with the shore : and although the 

 change in the coast-line trend may alter some aspects of the current, it seems to have a 

 relatively minor influence upon the degree of upwelling. 



EFFECT OF SEA-BOTTOM CONTOUR 



McEwen (1916), describing the horizontal distribution of temperature along the 

 west coast of North America, writes : " Upwelling of cold bottom water appears to be the 

 only type of circulation that could produce such a distribution of temperatures. Further- 

 more, the contour maps (pis. 1-3) reveal a striking correlation between the location of 



D.XIII 13 



