87 



evidence must be considered. If some deeper level, such as the 3,500 

 decibar surface, is the most nearly motionless then the direction of 

 flow at 2,000 decibars is opposite to that indicated by the arrows on 

 figure 41. In either case the direction of flow will be along the dynamic 

 isobaths shown. The temperature — salinity correlation was studied 

 at each end of the area in question. All of the observations from 

 depths greater than 1,500 meters from stations 2523 to 2528 east of 

 Flemish Cap have been plotted in figure 42 and the individual station 

 curves drawn. The two broken lines, separated by about 0.03°/°° 

 salinity, represent approximately the range of observational error. 

 A similar plot of temperature against salinity has been made in figure 

 43 for all observations from depths greater than 1,500 meters at sta- 



STATIONS 2523-2528 



STATIONS 2475- 



FiGURE 42.— Temperature-Sijlinity Figure 43— Temperature-salinity 

 correlation for all observations correlation tor all observations 

 deeper than 1,500 meters at sta- deeper than 1,500 meters at sta- 

 tions 2523-2528 occupied July 7-9, tions 2475-2484 occupied June 22- 

 1937, east of Flemish Cap. 25, 1937, southeast of the Tail of 



the Grand Banks. 



FiGURK 44.— Comparison of the 

 temperature-salinity correlation 

 for depths greater than 1,500 me- 

 ters below the Atlantic Current 

 east of Flemish Cap and south- 

 east of the Tail of the Grand 

 Banks. 



tions 2475 to 2484 on the section running southeasterly from the Tail 

 of the Grand Banks. Again the individual station curves have been 

 drawn and the envelopes indicated by broken lines. The separation 

 of the envelopes is again about the range of observational error except 

 in the vicinity of about 2.75°, where the characteristic shape of the indi- 

 vidual station curves require a lower-salinity hump. From figures 42 

 and 43 have been derived the mean curves shown in figure 44 and 

 representing the temperature-sahnity correlation for the two extreme 

 edges of the survey, along the course of the current, for depths greater 

 than 1,500 meters. From figure 44 it will be seen that the two corre- 

 lation curves cross and that in the upper part of the figure the water 

 off Flemish Cap is fresher than off the Tail of the Banks whereas in 



81041—39 7 



