and the individual scales had widths of 2 ppt starting at 30 o/oo and 

 increasing 1.5 o/oo per scale step through six scales. Beginning at 

 Station 53 the potentiometer of the potentiometer recorder was shunted 

 with a precision resistor to make it span 4 o/oo on each scale. There- 

 after there seldom was need to use any but Scale 1 (30-40 o/oo) , ex- 

 cept when the shunt on the conductivity cell was used. The temperature 

 range -2 to + 8 C was used exclusively. 



The instrument was subjected to an unusual amount of electri- 

 cal interference from the search radar or some of the equipment involved 

 in helicopter launches. Useful data could not be obtained at a number 

 of stations because of this factor. At these times the instrument usually 

 was replaced by the Beckman conductivity-temperature meter. 



The effect of the shunt was determined by finding the apparent 

 conductance change it produced in the STD in air. Tables of seawater 

 conductivity were then used to construct a table of equivalent salinity 

 change as a function of salinity and temperature. Over the limited tem- 

 perature range involved, this turns out to be a simple second degree 

 equation. A controlled calibration in seawater was not possible so a 

 check of the equation was attempted by comparing salinities where the 

 shallow and deep lowerings overlapped. This did not work well because 

 the depth of overlap was in a region of sharp temperature and salinity 

 gradients and the depth errors of the STD created substantial uncertain- 

 ties. There was a rapid change with time in this depth zone. For these 

 reasons it has not been possible to certify that the computed salinities 

 are any more accurate than ±0.3 o/oo. 



Similar problems are involved in the Beckman conductivity- 

 temperature meter. Its results were compared with the STD at points of 

 overlap where conditions appeared fairly stable. The mean temperature 

 error was -0.29°C and the salinity error -0.41 o/oo. The standard de- 

 viations of the mean were 0.04° and 0.06 o/oo respectively. The mean 

 corrections were applied but considerable uncertainty remains, particu- 

 larly because it could not be determined if there were non-random de- 

 viations . 



It should be noted that temporal and spatial variations in 

 water properties are so great in the region in and above the thermocline, 

 where the shunt and the RS5 were used, that errors of the magnitude 

 cited above are unimportant. 



The STD itself was continually standardized by means of a 

 Nansen bottle, just above the instrument, which was tripped at the 

 greatest depth of lowering. The salinity error averaged 0.09 o/oo with a 

 standard deviation of the mean cr of 0.01 o/oo. The temperature error 



