the Thermodynamic Scale of Temperature. 



139 



however, for the purpose of calculating this small term, will 

 be the value of t found by means of the uncorrected readings 

 of the gas-thermometer. 



The following table shows the corrections required for the 

 constant-volume thermometer in the case of the gases 

 hydrogen and nitrogen. The corrections, are given for every 

 10 degrees between the freezing-point and the boiling-point 

 of water, and the initial pressure at the freezing-point is 

 supposed to be 1 atmosphere. 



Tem- 

 perature. 



Correction for 



Hydrogen 

 Thermometer. 



Correction for 



Nitrogen 

 Thermometer. 



Tem- 

 perature. 



Correction for 



Hydrogen 

 Thermometer. 



Correction for 



Nitrogen 

 Thermometer. 



0°C. 











60° C. 



-•0007 



-•0025 



10 



-0003 



-0011 



70 



-•0006 



-•0021 



20 



-•0005 



-•00L9 



80 



-0004 



- 0016 



30 



- -0007 



-•0021 



90 



- 0002 



-•0009 



40 



-•0007 



-•0021) 



100 











50 



-0007 



-•0027 









It will be seen that the correction is very small in the case 

 of both gases; and it seems quite likely that the correction is 

 really smaller than the uncertainties due to errors of experi- 

 ment. We have already seen that the estimates for t derived 

 from the hydrogen and the nitrogen thermometers respectivelv 

 differ by as much as 0°'2 C. The source of this difference 

 cannot be located with certainty, and it is possibly due to the 

 portion of the numerical data contributed by M. Chappuis. 

 Any one who has read through M. Chappuis's paper will be 

 struck with the great care and skill he has shown ; but there 

 may still be unknown sources of error affecting his work, and 

 mere repetition of the same experiment cannot lead to 

 the detection of an error inherent to the method. If we 

 imagine that some such unknown source of error exists, and 

 also that it is somewhat irregular in its variation with 

 temperature, then mistakes comparable with, though less than 

 o, 2 C. might be made in the readings of the gas-thermometer 

 at temperatures between the freezing-point and the boilino-- 

 point. Mistakes due to such a cause might easily overpower 

 the small thermodynamic correction we have investigated 

 above. 



A careful comparison of the readings given by the nitrogen 

 and the hydrogen thermometers respectively was undertaken 

 by M. Chappuis (Joe, cit.). He employed the constant- 



