partly b}" forces associated with the North Atlantic eddy. It was 

 assumed that the area would be increased by 10,000 square kilometers 

 for each million cubic meters per second volume of flow of the Labra- 

 dor Current entering the area past section U. After subtracting such 

 an amount from the area to obtain an adjusted area, A, this adjusted 

 area was considered to represent the effects of causes associated with 

 the Gulf Stream S3^stem. Assuming that these causes were related 

 to the difference in sea level between Bermuda and Charleston, S. C, 

 the fluctuations in A would be related to the fluctuations in this sea 

 level difference. If sea level at Charleston minus the departure from 

 average sea level at Bermuda, expressed in feet, is H and A, expressed 

 in units of 10,000 square kilometers, is the adjusted area ISK months 

 later, 27 surveys made during the ice patrol seasons 1934-41 give the 

 expression 



A=6.8 (H-5.07) + 1.34 

 No surveys were made during the years 1942-47. Ten surveys 

 made during 1948-52 did not agree with the above relationship but 

 followed approximately the same relationship if a time lag of llK 

 months were used. Combining all 37 surveys and using the phase 

 difference of 13}^ months for the prewar data and llK months for the 

 postwar data the expression became 



A=6.97(H-5.07) + 1.67 



However, the correlation was much poorer and it is evident that the 

 full explanation involves more than has so far been considered. 



In 1953, the first survey gave a gross area of 9.63 with an adjusted 

 area A of 6.67 and in the fourth survey the gross area measured 6.95 

 for an adjusted area of 3.22, both of which are widely different from 

 the values of A computed with either of the above formulae. The 

 phase difference could have been either 11}^ months or 13K months 

 during 1953. 



The postwar da,ta now include 12 surveys with 4 each made during 

 April, May and June. The means of these actual values of A plotted 

 on the curve of postwar averages of the Charleston minus Bermuda 

 departure show the following possible time lag correlations: 1 month 

 negative; 2)^ months positive; 4 months negative; 6 months positive; 

 9K months negative; 11^2 months positive; 13 months negative; and 

 14)2 months positive. Examining each year, survey by survey, shows 

 only one of these (llK months positive) to have no years where the 

 sign of the correlation was contradictory. The quantity (H-5.07) 

 has continued to have low values whereas the actual value of A has 

 remained high. A decreasing value of (H-5.07) is interpreted to mean 

 an increasing activity in the North Atlantic eddy. If the adjusted 

 area A remains high it may be because of an anomalous failure of the 

 Labrador Current to increase. Another possible interpretation is 

 that the Atlantic Current water has decreased in salinity in its T-S 



80 



