CORRECTION OF THERMOMETERS AND DETERMINATION OF DEPTH OF REVERSAL 



21 



scaled for salinity and temperature at selected depths 

 and values of cr^j^ computed for these depths. The values 

 of (TtD so derived may then be used to construct a more 

 accurate density-depth curve which can be used to check 

 the values of mean density used in the reduction of the 

 thermometer depths. If the values do not check within 

 the limits mentioned above, a second approximation 

 must be made, but this will rarely be necessary. 



From the foregoing it will be seen that one meter in 

 depth corresponds to a difference of about 0.°01 C be- 

 tween the corrected readings of the protected and unpro- 

 tected thermometers. Experience has shown that unpro- 



tected thermometers having a range of about 60° C 

 divided into 1/5' can be read with an accuracy of better 

 than 0.°01. Comparisons of thermometer depths with 

 depths determined by wire length when the wire angle 

 was small indicate that the method gives depths reliable 

 to within about +10 meters. The use of unprotected ther- 

 mometers at intervals along the length of a wire to which 

 a number of water bottles is attached, in conjunction 

 with meter-wheel readings, thus provides a satisfactory 

 method of determining the depths of all the water bottles 

 on the wire. 



LITERATURE CITED 



Hidaka, K. 1932. Mem. Imp. Marine Obs. Vol. 5, no. 1, 

 p. 11. Kobe, Japan. 



Schumacher, A. 1923. Ann. Hydrogr. Vol. 51, p. 273. 



