Water samples were collected with tin-lined Nansen bottles. Each 

 bottle was equipped with two protected reversing thermometers, and 

 some were equipped with an additional unprotected reversing thermometer 

 in order to obtain thermometric depths. Precision of the protected 

 thermometers was good (Table I) , and it appears that errors of more 

 than +0.02°C were rare. Occasionally, the paired thermometers were 

 rearranged to reduce the possibility of recurring systematic errors. 

 Thermometric depths were computed by the L-Z curve method described 

 in H.O. Pub. 607. Often, agreement between unprotected thermometers 

 in a cast was not good, and it was necessary to consider the thermometers' 

 previous histories when constructing the curves. 



TABLE I. Temperature Differences Between Pairs of Protected Reversing Thermometers 



Temp. 



Diff. 



°C 



0.00 



0.01 



0.02 



0.03 



0.04 



0.05 



0.06 



Only 



One 



Accepted 



None 

 Accepted 



No. 



Obs. 



225 



268 



115 



47 



7 



3 



1 



5 



6 



Salinities were determined aboard ship using an Industrial Instruments 

 inductively coupled salinometer (Model RS-7A) . Although a precision 

 of +0.003% is possible with this instrument in the laboratory, 

 the salinity determinations probably were less precise under the 

 more rigorous conditions which existed on the EDISTO. To assess 

 the accuracy of the salinity determinations, an examination of the 

 salinities of all samples (97) from below 500 meters with temperatures 

 less than 0°C was made. Data presented by Gladfelter 1964, Helland- 

 Hansen and Nansen 1909 and 1912, Mosby 1959, Nansen 1915, and Sverdrup 

 1933 indicate that these samples should have a narrow salinity range 

 with a mean salinity of approximately 34.91 to 34.92 %, . Excluding 

 questionable values, the average salinity of the 97 EDISTO samples 

 is 34.91% • If it is assumed that the 'true' salinity variation 

 in these samples is +0.01% » then 90% of the observations fell 

 within +0.02C of the 'true' range, 94% fell within +0.03%,, , and, 

 except for one value, all of the salinities fell within +0.04^ of 

 the 'true' range. Values which were considered questionable were 

 included in this comparison. 



Dissolved oxygen and nitrogen concentrations were determined by 

 the gas chromatographic method described by Swinnerton and Sullivan 

 (1962) . It was assumed that the dissolved nitrogen concentrations 

 of the sea water samples should approximate the 100% saturation values 

 given by Rakestraw and Emmel (1938) , and when the deviations from 

 saturation were greater than +10%, the results were not accepted. 

 Ninety-nine comparison samples were analyzed for dissolved oxygen 

 by the Naval Oceanographic Office's (NAVOCEANO) modification of the 

 Winkler titration. However, comparisons were made only when the 



