300 m. or less, and at that level velocities of this magnitude were found only 

 between stations 29 and 30. In contrast, during; cruise 10 velocities of over 20 

 cm/sec (0.4 kt.) were found on the 300-decibar surface. 



On the windward side of the islands the dynamic topography contours are 

 Incomplete because of the omission of the stations north of Hawaii. The flow 

 normal to the section east of Hawaii was confused in the upper 150 m. by the east- 

 erly flow indicated between stations 28 and 29. North of Oahu the flow was again 

 northerly, indicating that the major portion of the flow was being deflected by 

 the middle group of islands. 



To the leeward of the islands the most prominent features were again the 

 two cyclonic eddies. The one west of Hawaii was larger in diameter, but did not 

 have velocities as great as the corresponding eddy of cruise 10. The eddy south 

 of Oahu was centered off station 1. It was also relatively weak, having a maximum 

 velocity of only 30 cm/sec {0.6 kt.) as against the 72 cm/sec (1.4 kt.) maximum of 

 cruise 10. 



In Kauai Channel northerly flow was indicated in the upper 100 m. It was 

 again the result of the high values at station 8, which were indicative of a clock- 

 wise flow around Oahu. 



GEK Currents 



The current vectors computed from GSK observations are shown in figure 25. 

 They represent POFI's first attempt at direct current measurements in the Hawaiian 

 area. The results are only partially satisfactory, if the surface geostrophic 

 currents are used as the standard. They agree fairly well in direction and magni- 

 tude in the offshore areas where the velocities were large, such as in the leeward 

 eddies, but in the vicinity of station 8, just west of Cahu, the G3K currents were 

 southerly and the geostrophic northerly. As mentioned above, the accuracy of the 

 latter are also open to doubt in inshore areas, but it is doubtful if the error 

 could be over 90°, as the GEK currents would indicate. 



The persistence of the eddy around station 1, shown by the "recheck" of 

 the area, was the most interesting information provided by the G£K. In spite of 

 time differences of up to 7 days and a shift of the incident winds from east to 

 south, the observations made on the second coverage of the area (those marked "A" 

 in fig. 25) agree almost perfectly with those made on the first. 



Temperatures 



Again the temperature field (figs. 26 and 27) was simply a reflection of 

 the distribution of mass associated with the circulation pattern shown by the geo- 

 strophic currents. Divergence and/or wind-mixing in the eddy west of Hawaii was 

 shown by the 2°F. drop in surface temperature at its center. 



Sigma-t 



The sigma-t distribution (figs. 28 and 29) differs only slightly from 

 that of cruise 10. On the windward side, east of stations 17 and 18, the 23.2 

 sigma-t surface is still the upper limit of the well-defined density gradient, but 

 over the rest of the area it is the 23.0 surface. The difference was caused by 

 slightly lower surface temperatures in the southeastern sector of the windward side. 

 This difference could be attributed directly to mixing or "upwelling" only in the 

 area around station 19, where' the 23. sigma-t isopleth reached the surface and 

 surface salinity was slightly greater than in the surrounding area. 



The horizontal plots of depth of sigma-t surfaces are again included for 

 the information which the 25.0 and 26.0 surfaces add to the description of the cir- 

 culation north of the middle group of islands. The cells of greater than 200 m. in 



14 



