actual area and from it the location of the boundary itself, there still 

 remains the prediction of the volume of flow of the Labrador Current 

 past section U, which volume of flow is the difl"erence between the 

 actual and adjusted areas. The absence of tide gauge data or even appro- 

 priate locations for tide gauges which might give the information sought, 

 upstream from the Grand Banks with respect to the Labrador Current, 

 reduces the possible sources of data to meteorological observations and 

 such scanty oceanographic observations as have been made. 



Gradients in barometric pressure anomalies seemed to be the most 

 promising of the data available and several gradients were examined. 

 Of those investigated, the best correspondence with yearly means of 

 observed volume of flow seemed to be a combination of the December 

 and January gradient between Belle Isle and Julianehaab with the 

 January gradient between Cape Race and Belle Isle. The former 

 gradient has a reasonable physical interpretation as being a measure of 

 the wind driving-force tending to increase the total Labrador Current, 

 and the latter gradient presumably is effective in determining the pro- 

 portions with which the total Labrador Current divides into that branch 

 which remains close to the Avalon Peninsula of Newfoundland, and 

 the eastern branch with which we are concerned which flows south 

 along the eastern edge of the Grand Banks. The best correlation sus- 

 piciously gives the latter gradient 10 times the weight of the former. 

 Conceivably the latter gradient is also a determining factor in the 

 volume of flow of the Grand Banks whorl which is a part of the volume 

 of flow measured at section U. 



In figure 17 the solid curve connects points representing the mean of 

 each season's observations of the volume of flow past section U for the 

 years 1934 to 1941. The line of alternate long and short dashes shows 

 the barometric pressure anomalies in the gradients Belle Isle minus 

 Julianehaab for December and January and Cape Race minus Belle 

 Isle for January with the latter being given 10 times the weight of the 

 former. It will be seen that the curve for volume of flow trends upward 

 and the curve for barometric pressure anomalies trends downward. 

 Evidently the pressure anomalies do not give the whole picture and 

 some important factor having a long period trend has been neglected. 

 Inasmuch as the Grand Banks whorl is considered to be the result of 

 differential vernal warming of the waters over the shoaler part of the 

 Banks, it is possible that such a factor as a long period increase in tem- 

 perature may be the cause. At the present it remains unknown and an 

 empirical formula of the following form is the best that can be ofi^ered 

 now: 



-(T) 



1.25 + 1.58 (A^-1934) 



where V is the average volume of flow (in cu.m/sec. x 10^) of the 

 Labrador Current past section U for the jjeriod April, May, and June 



23 



