salinity anomaly, is really only a criterion of water characteristics, rather 

 than one of movement of water; that the southward shift of this apparent 

 boundary represents colder fresher water (perhaps slope water) drawn 

 into the North Atlantic eddy circulation from peripheral waters in the 

 vicinity of the Charleston- Bermuda section; that the northward move- 

 ment of the apparent boundary represents warmer, saltier Mediter- 

 ranean water drawn into the circulation of the eddy from peripheral 

 waters in the southeastern sector; and that 53^^ months and 133^ months 

 represent the travel times for these peripheral waters between their 

 junction points and the Grand Banks region. Such an explanation 

 would require the North Atlantic eddy to act somewhat as a rigid 

 wheel with all parts or at least the western and southeastern parts 

 acting together in fluctuations in volume of flow, and with time lags 

 existing only in water characteristics and not in transport. 5 It would 

 be a convenient explanation of the semiannual cycle in volume of 

 flow with westerlies and trades as the driving forces applied alternately 

 with an ineffective period between each. 



However, it can be tested by determining whether the seasonal 

 fluctuation in difference in sea level across one sector of the eddy is 

 in phase with the fluctuation in another sector. The difference in sea 

 level, as determined by tide gauges and corrected for barometric 

 pressure, between Horta in the Azores and Lagos, Portugal, has been 

 examined and monthly normals based on the interval from 1927-36 

 have been determined. A smooth curve drawn through these mean 

 monthly mean values shows the seasonal fluctuation in difference in sea 



^ In "Technical Report No. l.S on the Hydrography of the Western Atlantic; Annual 

 Variations in Current Speeds in the Gulf Stream System," an unpublished report of 

 the Woods Hole Oceanographic Institution to the Oceanographic Division of the 

 Hydrographic Ofifice, U. S. Navy, November, 1948, Frederick C. Fuglister presents 

 curves and tables showing the seasonal variations in average surface current speeds 

 derived from ship reports for several areas in the western half of the North Atlantic 

 eddy. His area 4 is located off the coast between Florida and Cape Hatteras and may 

 be taken as corresponding to the Charleston-Bermuda area. His area 6 is located 

 south-southwest of the Tail of the Grand Banks and may be taken as corresponding 

 to the Grand Banks area. His area C is located north-northwest of the Azores and 

 corresponds to an area from which no other data are discussed in this Bulletin. 



The data used by Fuglister do not in themselves permit the deduction of anything 

 but seasonal normals and he states that the time of year of maximum current speed 

 varies as we move around the system, occurring in summer in the southern portions 

 and in winter in the northern portions; but the minimum speeds occur during the fall 

 months in all areas presented, and that all areas show secondary maxima and minima. 

 His data, in general, show two maxima and two minima which occur in the three 

 areas mentioned at the following times of year: 



Maxima Minima 



Area 4 (Charleston) — Jan.; June-Aug. Feb.; Nov. 



Time lag 14 13 16 12 Av. time lag 13 Ji months 



Area 6 (Grand Banks) Mar.; July-Sept. June; Oct.-Dec. 



Time lag 9}4 10 9 lOJ^ Av. time lag 9M months 



Area C (Azores) — Dec-Jan.; June Mar.; Sept.-Oct. 



Thus there seems to be some support here for the time lag of 133^ months, derived 

 by other means, between the Charleston-Bermuda area and the Grand Banks area; 

 and to suggest that the time lag between the Grand Banks area and that north-north- 

 west of the Azores may be about 10 months. These time lags, in the light of the material 

 presented in this Bulletin, would be interpreted as time lags in the latitudinal shift of 

 the northern boundaries of the North .Atlantic eddy at the Grand Banks area and in the 

 vicinity of- the Azores, rather than in total volumes of flow of the eddy at these points. 



20 



