the meter wheel readings and surface wire angles. 

 A roughly parabolic curve was assumed for the 

 wire, this being approximated by four segments 

 each liaving a progressively snuxller inclination 

 from tlie vertical with increasing depth. The 

 uppermost segment was assigned the surface wire 

 angle, while the lowermost was assumed vertical. 

 The two segments in between were assigned in- 

 termediate angles. A comparison between this 

 depth calculation method and that using unpro- 

 tected thermometers in cases where these meas- 

 urements were a\ailable suggests that depths so 

 calculated are accurate to within about ±'25 m 

 for the most extreme wire angles encountered 

 (nearly f!<t°). Foi- the majority of the calculated 

 depths the error is felt to be considerably smaller. 



All salinity samples were analyzed aboard ship 

 using a Bissett-Berman Hytech Model 62'20 

 Portable Laboratory Salinometer supplied by 

 the Coast Guard. A standard was run about 

 every second station, or more frequently when 

 felt necessary, using Copenhagen Water. A me- 

 chanical malfunction in the salinometer increased 

 the time required to analyze a sample from 2-3 

 minutes to as much as 10 minutes, but did not 

 affect its accuracy. The advertised accuracy of 

 this salinometer is ±0.(K)3°/oo, but considering 

 temperature compensation errors, power supply 

 fluctuations and possible operator error it is felt 

 that ±0.02°/o„ would be a more realistic value. 



Samples intended for phosphate, nitrate and 

 silicate analyses, taken at all stations and depths 

 on the latitudinal sections, were decanted into 2 

 oz. polyetiiylene bottles, tightly capped and im- 

 mediately frozen. After freezing, tlie bottle caps 

 were retightened to prevent sublimation of water 

 from the sam])le or, con^•ersely, fresh condensa- 

 tion of ice witliin tlie sample bottle. The nutrient 

 analyses were carried out on a Technicon Auto- 

 Analyzer at the University of Washington's De- 

 partment of Oceanography. Accuracy of these 

 determinations depends upon the relative con- 

 centi'ation of the nutrient in tlie sample, being 

 about ±1% in general. If concentrations are very 

 small, as in the case of phosphate, the error may 

 be as large as ±2%, and for larger concentrations 

 the error will be correspondingly less (Mr. Roy 

 Olund, Department of Oceanography, University 

 of Washington, personnel connnunication). 



Cui'rents were recorded on two different makes 

 of meter; Braincon Corporation Type 316 Histo- 

 gram Current Meters were used at the 100m 

 deptli at each location and Geodyne Model 850-1 



meters were used at 300m. The former instru- 

 ments were set to record over 15 minute intervals. 

 Tlie recording modes of the two instruments were 

 basically different. The Braincon instruments re- 

 corded over nine minute intervals, using the tenth 

 minute to cycle to the next frame, producing an 

 analog record on 16 mm film and yielding essen- 

 tially the averaged current (speed and direction) 

 over each nine minute interval. The Geodyne 

 instruments recorded instantaneously during a 

 pre-set short interval once every 15 minutes on 

 magnetic tape, so that essentially instantaneous 

 rather than average values were obtained. 



The effect of tlie different recording modes on 

 the measurements is uncertain, althougli spectral 

 analyses applied to these data might yield dif- 

 ferent results for the higher frequency spectra 

 depending on which recording mode was used. 

 Further research is currently being planned at 

 the University of Washington to detennine the 

 effect of the recording mode upon the measure- 

 ments, and the relations between the measured 

 and true currents in regions of rapid temporal 

 variation. 



Accuracy of the speed measurements for the 

 Braincon instruments is estimated at ±3% of 

 the measured values. Experience in the Bering 

 Sea and on ice island T-3 has shown that direc- 

 tion measurements made at high latitudes with 

 the Braincon meters are seldom accurate to more 

 than ±20° due to weak horizontal magnetic 

 fields (Mr. Richard B. Tripp, personal communi- 

 cation). The accuracy of the Geodyne instru- 

 ments is also estimated as ±3% of the measured 

 value. Little experience is available concerning 

 the directional response of the Geodyne meters 

 at high latitudes, but the mode of construction 

 of tlie direction densing device suggests that they 

 are comparable to tlie Braincons, or accurate to 

 about ±20°. 



The method of mooring the meters is depicted 

 in figure 2. A 700 lb. railroad wheel was used as 

 file anchor, and the mooring line was %" braided 

 nylon, the braided form being used to minimize 

 line twist. The arrays were moored by first 

 streaming the entire system forward from the 

 foredeck wliile steaming slowly astern. The sur- 

 face float was therefore placed in the water first 

 and was furthest from the ship, while the anchor 

 was di-opped last. The arrays were reco\ered by 

 hauling in the whole system by means of the 

 anchor windlass. 



AYhen retrieval of the current meters was be- 



