mixed water for 1951 the abnormal mixed water found during 

 the first survey was excluded from consideration. An approxi- 

 mate depth scale in meters is shown on each curve in figure 11. 

 It will be seen from figure 11 that, level for level, each water mass 

 was lighter during 1951 than the 8-year mean and that lower 

 salinities were responsible for these differences in densities. The 

 diff'erences were greater in the upper levels than in the deeper 

 levels, but they still averaged 0.02 to 0.03 in Ot between 400 and 

 1,000 meters. 



The second survey was not included in the analysis of the T-S 

 relationships of the water masses of the Grand Banks region 

 inasmuch as the data considered previously were from the area 

 south of Flemish Cap. The stations lying seaward of the Lab- 

 rador CuiTent in the second survey were examined to find out 

 whether they approximated the characteristics of the mixed water 

 of the Grand Banks region. Only stations 4440 to 4445 and 4459 

 to 4464 did. These 12 stations were located in the vicinity of 

 Flemish Cap and lie in the region where the dynamic topography 

 indicates the water entered the area from the Grand Banks region 

 farther to the south. It is considered probable that the charac- 

 teristic mixed water of the Grand Banks region is not formed 

 northward of Flemish Cap. 



In comparing the results of the von Arx current meter (geo- 

 magnetic electrokinetograph, abbreviated GEK) with the currents 

 deduced from density distribution a method was presented in Bul- 

 letin No, 34 of this series wherein the currents as measured by 

 GEK during the run between two oceanographic stations were 

 converted into an equivalent difference in dynamic height at the 

 two stations using the expression AD ~ 0.01391 LC sin ^m where 

 AZ) is the difference in dynamic height in dynamic meters between 

 two points separated by a distance L nautical miles, C is the aver- 

 age component of the current normal to the line between stations 

 expressed in nautical miles per hour and ^m is the mean latitude. 



If C is not zero there is an angular difference between the ship's 

 heading and the course made good. The jog component Mj meas- 

 ured by GEK is normal to the ship's heading. We require the 

 component C normal to the course made good. L and <^,n are known. 

 Let the mean component of the current in the direction of the ship's 

 heading be Mc (derived from the values computed each half hour) . 

 Let the time elapsed during the run be t hours. The mean Mj can 

 be derived from scaling the trace at short intervals of time. The 

 total movement of the ship from the current acting during t is sub- 

 tracted vectorially from L at the station of arrival to reach the dead 



32 



