CIRCULATION NEAR THE WASHINGTON COAST 591 



meters of depth were occupied at intervals of 2 to 12 hours in the 

 several experiments in an attempt to correlate the results with the 

 dynamic heights. Rotary changes in current direction were readily 

 demonstrated, but twenty-four-hour periods are too short for effective 

 harmonic analysis. Two experiments of 3 to 4 days' duration were 

 therefore made in June and August 1953, Cruises No. 29 and 31, re- 

 spectively. 



In Cruise No. 29, the ship steamed backward and forward two 

 hours' run on reciprocal headings, taking G.E.K. fixes every hour to- 

 gether with other data. Positions were fixed frequently by loran, and 

 hydrographic stations were occupied about every 12 hours. Some fail- 

 ures in equipment being used for corollary measurements caused several 

 undesirably large gaps in the data, but the results are only a little less 

 consistent than those to be presented below. 



In Cruise No. 31 the cruising plan was modified, as suggested by 

 Mr. Joseph Reid of Scripps Institution of Oceanography, so that the 

 ship steamed squares on the cardinal compass headings, each side being 

 approximately 7 minutes' run. Thus a fix is obtained every 7 minutes, 

 there being some interdependence in the measurements because each 

 datum enters into two fixes and two zero determinations. Reid has 

 directed experiments of this type in which he has simultaneously fol- 

 lowed a freely drifting buoy (personal communication). An attempt 

 to do this with an improvised buoy and drag was abandoned after the 

 buoy lost its drag and was itself nearly lost at night during a radar 

 failure. 



The results of this experiment have been expressed as north and 

 east components of velocity and are presented in Figure 10. K has been 

 assumed to be unity and corrections have been made for electrode 

 droop. The results are surprisingly consistent and have given rise to 

 a renewed belief in the reality of G.E.K. measurements. The 16-hour 

 inertial period is evident by visual inspection. Much of the distortion 

 is due to other components, principally the semidiurnal. Some smooth- 

 ing has been practiced but there is difficulty in deciding which fluctua- 

 tions may be real and which due to experimental uncertainty. Due to 

 the interdependence of separate fixes, errors often appear symmetrically 

 in alternate or adjacent points. Moreover, in some cases the record 

 can not be interpreted more accurately than several tenths of a millivolt. 



In correlation with this experiment, an anchor station was occupied 

 for the preceding three and one-half days atop Cobb Seamount (see 

 Fig. 8). Here currents were measured at a depth of 20 meters with 

 an Ekman current meter every 30 minutes, and temperature structures 

 were measured hourly by bathythermograph. Bathythermograms were 



