'f^.â^.40. 



50 

 100 



Stat. h la. ho W 



Fig- 55- Curves showing- the Vertical Distribution of Density at 

 Slats. 40 and 41 a (and 41!. 



At the Surface 11. 6: sin 66" = 12.5 cm. /sec. towards S 66" W 



10 metres 0.6: sin 44*^= 0.9 

 3.6: sin 55"= 4.4 



2.3: sin 38' 



3-7 



N44OE 

 N55OE 

 N38O E 



These values of velocity are so near those found by 

 our current-measurements at the same depths that the agree- 

 ment is probably more or less accidental. 



At 20 metres the average current should have been 

 directed towards Nô'^W (with a velocity of 7.9 cm. /sec.) 

 according to our current-measurements. This is very nearly 

 along the direction of our section, and the current at this 

 depth should thus have had a quite insignificant component 

 of 0.8 cm. 'sec. only) going in the opposite direction of that 

 found by our computation. In this case, therefore, the com- 

 puted value (7.2 cm. /sec.) does not agree with that found 

 by the measurements (7.9 cm. /sec), on account of the observed average 

 direction of the current. The latter was only observed at Stat. 41, and we 

 do not know what the direction may have been in the region about Stat. 40. but 

 our current-chart, Fig. 66, based on the distribution of density, indicates that 

 it should have been approximately the same. It has, however, also to be 

 kept in view that, at this depth, very slight vertical oscillations of the strata 

 at Stats. 40 or 41 a, might have been sufficient to alter the densities 

 essentially. 



On the other hand, it is a general feature in sections going in the 

 same direction as a current, that the densities are heavier at the stations 

 towards which the current moves, than at those from which it moves. Such 

 sections will therefore, if directly used for computations of the velocities 

 in the above manner, give erroneous results. This longitudinal distribution 

 of the densities, in the currents, will naturally influence the inclinations of 

 the isopycnals in all sections that are not directed perpendicularly to the 

 direction of the current. This circumstance should, therefore, also be taken 

 into consideration by the computations of the velocities from such oblique 

 dynamic sections, in order to improve the accuracy of the results. 



