the lower part of the figure, in the deeper water, that off the Tail of 

 the Banks is fresher than that off Flemish Cap. If we consider that 

 the source of this fresher water of more negative salinity anomal}- is 

 the downward mixing, cabbeling, mixed water containing Labrador 

 Current water as a constituent and found all along the boundary be- 

 tween the Labrador and Atlantic Currents, then the interpretation 

 is that in the upper levels the Atlantic Current is traveling from the 

 Tail of the Banks toward Flemish Cap, and that in the deeper levels 

 the water below the Atlantic Current is moving in the opposite direc- 

 tion, from Flemish Cap toward the Tail of tbe Grand Banks. 



The corollary is that, at some level in between, the water is motion- 

 less and that the motionless level is at about where the two curves 

 shown in figure 44 cross. The presence of the hump in the curve for 

 the Tail of the Grand Banks makes it difficult to say from figure 44 

 where this motionless level is located more exactly than that it is 

 between temperatures of about 2.7° and 3.4° which correspond to 

 depths of about 1,750 and 3,200 meters respectively. Thus there is 

 strong evidence in favor of the existence of a motionless surface such 

 as was assumed in the construction of figure 4L Also, the assumed 

 depth of this surface is within the limits indicated above from con- 

 sideration of figure 44. 



Another check on the assumption that the 2,000 decibar surface is 

 most nearly motionless is made possible by the distribution of the 

 isobaths in figure 41 with respect to the overlying Atlantic Current. 

 The volume of flow of the Atlantic Current past the section east of 

 Flemish Cap, determined by numerical computation and checked by 

 planimeter measurements of a velocity profile, was 39.3 million cubic 

 meters per second relative to the 2,000-decibar surface. Figure 41 

 indicates that the deep water velocity is greater at this section than 

 in any other part of the survey. It should be mentioned here that 

 accumulation of errors make the deep water velocity values less reliable) 

 than the values nearer the surface. However, it is believed that they 

 are sufficiently accurate to furnish a significant check as to the direc- 

 tion of motion of the water between the 2,000 and 3,500-decibar 

 surfaces. If, as was assumed in the construction of figure 41, the 

 2,000-decibar surface is motionless, the average velocity at the 3,500- 

 decibar surface between stations 2523 and 2527 is 6.6 centimeters per 

 second southerly and the average velocity of the water in the rectangu- 

 lar cross section between these two levels and these two stations is 3.3 

 centimeters per second southerly. However, if the 3,500-decibar 

 surface is motionless the average velocity between these two stationsi 

 is 6.6 centimeters per second northerly at the 2,000-decibar surface 

 and the average velocity of the water in the previously mentioned 

 rectangular cross section is 3.3 centimeters per second northerly and 

 all of the velocities used in determining; the volume of flow of the 



