an accuracy of approximately ±0.5%o of chlorinity (±0.93% of salinity). 

 It also has been concluded by CBI from 679 simultaneous measure- 

 ments by two such instruments that the standard deviation of measure- 

 ment errors for the CTI is not greater than 0.05% of chlorinity (0.09 %o 

 of salinity). 



3. Temperature- chlorinity- depth recorder 



Messrs. B. V. Hamon and N. L. Brown have described a temper- 

 ature- chlorinity- depth recorder for use in the upper 1,000 meters of 

 the sea. This instrument measures temperatures from 0° to 30°C with 

 an accuracy of ±0.15 C C. The chlorinity range spans about 7%o of chlo- 

 rinity (28.9 %o to 41.6 %o of salinity), and the stated accuracy is ±0.05%o 

 of salinity if a surface water sample is taken at each station as a check 

 for cell drift. The accuracy in depth is ±10 meters. 



The three quantities are recorded sequentially by a single-pen strip 

 chart recorder, each quantity being recorded for five seconds at a time. 

 The underwater unit is suspended by means of an armored steel cable 

 which has a single electrically insulated core. This core carries power 

 down to the underwater unit and brings up the signals from the measuring 

 elements. A simple frequency modulation method of telemetering is 

 used. 



Chlorinity is measured by means of a conductivity cell with plat- 

 inized electrodes. The effect of temperature on conductivity is com- 

 pensated by using a thermistor. Over the temperature range 5° to 

 25°C, compensation to within ±0.02%o of salinity can be attained. At 

 0°C and 30°C, however, the correction is about plus 0.2 %o and minus 

 0.2 %o of salinity, respectively. 



It was found necessary to avoid direct electrical connection between 

 the conductivity cell lead and the pressure vessel, as such a connection 

 allowed the flow of electrolytic currents that caused rapid polarization 

 of the electrodes. The oscillator circuit and its power supply are, 

 therefore, insulated from the pressure vessel for direct current, but 

 effectively grounded for alternating current by a capacitor. 



The equipment was used at sea about 60 times between April 1955 

 and November 1957. Nansen bottles and reversing thermometers were 

 used as checks. The difficulty in comparing results was caused mainly 

 by errors in depth. Correction for the effect of pressure on conductivity 

 must be applied and a further correction for temperature. 



IV- 5 



