The persistence of these anticyclonic eddies 

 and a large percentage of the other eddies 

 down to the 400-, 600-, and even to the 800- 

 decibar surfaces, as shown by the cross sec- 

 tions, is further indication that they are part 

 of the general circulation and not due to local 

 or transient phenomenon such as wind shifts 

 and internal waves. 



A comparison of the dynamic 

 heights of the stations which were occupied 

 close together in space but at widely different 

 times demonstrates that the assumption of 

 steady state for the cruise period gives a re- 

 alistic concept of the general circulation. The 

 heights of all surfaces at station 85 were lower 

 than those of station 55, which was occupied 

 only 2 miles away 17 days earlier. The cross 

 sections show, however, that the use of the 

 heights of either, instead of the average values, 

 would have merely increased (using station 85) 

 or decreased (using station 55) the velocities 

 around the south side of the anticyclonic eddy. 

 Figure 9 also shows that although the difference 

 in dynamic heights between stations 31 and 89, 

 which were occupied 33 miles apart in space 

 and 34 days in time, was as much as 0. 145 dy- 

 namic meters (at the surface), the topography 

 would have been basically the same if either 

 had been used instead of the average. 



The geostrophic current velocities 

 indicated by the spacing of the dynamic height 

 contours show that the flow was very weak even 

 in the well-developed eddies. The maximum 

 velocities at the surface occur in the three 

 eddies centered on 145 W. and are only about 

 0.4 knot (20 cm. /sec. ). The maximum velo- 

 city in the general zonal (easterly) flow at the 

 surface is about 0. 2 knot (10 cm. /sec. ). How- 

 ever, velocities of this magnitude only occur 

 in narrow bands and the average velocity would 

 be more in the order of 0. 1 knot (5 cm. /sec, ) 

 or less. On the subsurface levels the velocities 

 fall off rapidly below 100 meters. The maxi- 

 mum velocities of the general zonal flow have 

 decreased to less than 0. 1 knot over the entire 

 area at the 200-decibar surface, and in the 

 eddies they are less than 0. 2 knot at 200 

 decibars and less than 0. 1 knot at 400 decibars . 

 On the 600-decibar surface the velocities are 

 less than 0. 02 knot (1 cm. /sec. ) in the general 

 cyclonic flow through the area and have a maxi- 

 mum of between 0. 04 and 0. 06 knots (2-3 cm. / 

 sec. ) in the remnants of the two anticyclonic 

 eddies on 145 W, These values are well be- 

 low the limits of accuracy to be expected from 

 the dynamic topography and show that the use 

 of a surface deeper than 1,000 decibars, such 



as the 2, 500-decibar surface used in the 

 Carnegie Report (Sverdrup et al. 1944-1945), as 

 the reference level would not have made an ap- 

 preciable difference in the dynamic topography 

 of the upper levels. 



Geomagnetic Electrokinetograph Currents 



Current measurements at 30-mile 

 intervals were made with the Geomagnetic 

 Electrokinetograph (von Arx 1950) during the 

 second part of Smith cruise 25. The k-factor 

 was taken as unity, and correction was made for 

 the droop of the electrodes following procedures 

 outlined by Knauss (personal communication, 

 see McGary 1955). Figure 22 gives the results 

 together with the dynamic topography of the sea 

 surface with respect to the 1 , 000-decibar level. 



It is immediately evident that the 

 GEK is measuring what may be considered a dif- 

 ferent part of the "spectrum" of water movement 

 than that indicated by dynamic computations; the 

 results are extremely variable in direction and 

 attain considerably greater magnitudes than the 

 geostrophic currents. Barnes and Paquette 

 (1954) describe a very similar situation off the 

 coast of Washington; the net (geostrophic) circu- 

 lation there is also relatively slow and motions 

 with tidal or inertial periods dominate the indi- 

 vidual GEK measurements. They found that 

 calculating the data as 48-hour running means 

 removed most of these short-period variations, 

 but the results seemed to agree more with the 

 local winds than with the indicated geostrophic 

 current. They concluded that the instrument was 

 measuring real, transient, wind-driven currents, 

 which are integrated into the longer-term net 

 flow indicated by the distribution of mass. 



Mr. JosephReid at Scripps Institution 

 of Oceanography (personal communication) found 

 that simple averaging of the GEK currents by 

 calendar days seemed to remove most of the 

 short-term variations from measurements taken 

 off the coast of California, and the averaged re- 

 sults indeed show excellent agreement with the 

 geostrophic currents. Reid subsequently averaged 

 the Smith 25 data but the results show little agree- 

 ment with the dynamic topography of the surface 

 relative to the 1, 000-decibar level (fig. 23) . 



Comparison of the GEK currents with 

 the local winds (fig. 5) shows no obvious relation- 

 ship. These results are not surprising, consider- 

 ing the complex nature of the geostrophic currents 

 in the area and the large variations in the wind, 

 which during much of the time was directed 

 against the geostrophic surface current. 



