THERMOMETRIC DETERMINATION OF DEPTH 

 Table 2. Number of cases where the difference lies between stated limits 



13 



wire length for the different pressure thermometers. 

 The tabic gives the number of the thermometer, the num- 

 ber of observations with this thermometer, the mean 

 difference between the wire length and thermometer 

 depth, the total range of these differences, and the mean 

 depth as obtained by thermometers. These data are en- 

 tered for wire angles 0° to 5° and 6° to 10°. 



An inspection of the table shows that the differences 

 in depth between the wire lengths and thermometer 

 depths as a rule are smaller than 5 meters if the wire 

 angle is from 0° to 5° and smaller than 10 meters if 

 the wire angle is from 6° to 10° (see table 2). 



Only in three instances the mean difference is so 

 great that an error in either the correction of the ther- 

 mometers or the pressure factor seems to have influ- 

 enced the mean difference. This applies to thermome- 

 ters 1696, 2993, and 2996. According to an inspection 

 of the single values it seems probable that the correc- 

 tions of these thermometers have changed since they 

 were determined at PTR and, since an error in the tem- 

 perature correction introduces an error which is inde- 

 pendent of depth, the simplest procedure is to apply a 

 constant correction to the depths which are computed on 

 the basis of the original temperature corrections. For 

 the depths derived by means of these thermometers the 

 following corrections have, therefore, been adopted: 



Depth by thermometer 1696: correction: -20 meters; 

 Depth by thermometer 2993: correction: -10 meters; 

 Depth by thermometer 2996: correction: + 10 meters. 



After application of these corrections we find the follow- 

 ing mean differences: 



Wire angle 0° to 5°. Wire length minus ther- 

 mometer depth: -1.9 meters (109 cases). 

 Wire angle 6° to 10°. Wire length minus ther- 

 mometer depth: +3.2 meters (137 cases). 



It is seen that the mean thermometer depth is slight- 

 ly greater than the mean wire length in case the wire 

 angle is between 0° and 5°. This result may be owing to 

 systematic errors in the corrections of the thermome- 



ter (a greater rise of the zero point of the unprotected 

 thermometers than of the protected thermometers would 

 introduce an error of this sign) or it may be owing to a 

 small systematic error of the meter wheel, used for 

 measuring the wire length. It is of greater interest to 

 state that the difference increases when the wire angle 

 increases as should be expected. 



Examining the total ranges of the differences we find 

 that these are much smaller than the possible ranges 

 which were estimated on the basis of. the sources of er- 

 rors. We found that these errors might lead to errors 

 in the depth between +20 and -20 meters for depths 

 smaller than 1000 meters, which means that the range of 

 the differences between the exact values and the meas- 

 ured values of the depth might amount to 40 meters. At 

 greater depths this range would be greater. When com- 

 paring the thermometer depth with wire length we have 

 furthermore to bear in mind that the reading of the me- 

 ter wheel may not indicate the exact wire length because 

 the wire may have slipped on the wheel and we must, 

 therefore, expect the ranges in table 1 to be greater than 

 the estimated ranges (p. 12) provided that the errors of 

 the thermometers are as great as supposed. From table 

 1, however, we find: 



Table 3. Number of cases in which the total range 



of the difference, wire length minus thermometer 



depth, lies between stated limits 



From this compilation it is seen that the ranges are 

 smaller than estimated, and this result leads to the con- 

 clusion that the accuracy of the temperature determina- 

 tions is greater than supposed. 



Grouping the differences and ranges according to the 

 depths to which the thermometers have been used, we 

 find the values which are entered in table 4. 



Table 4. Differences between wire length and thermometer depth, and total 

 ranges of these differences 



