38 



REPORT — 1901. 



considers this impossible, and has recently ^ proposed an emended curve, 

 which would alter the correction by nearly one part in 1,000 at 100°. 

 (2) The experiments of Chappuis refer to the constant-volume nitrogen- 

 thermometer at one metre of mercury initial pressure, whereas the 

 difference-formula is assumed to refer to the constant-pressure air-ther- 

 mometer at 76 cm. pressure. The correction in the latter case is quite 

 different, so that we should not assume Chappuis' results for the reduction. 

 On the whole we shall probably be nearest the truth if we calculate the 

 correction for the scale of the constant-pressure air-thermometer from 

 the observations of Joule and Thomson ^ by the method which I have 

 explained in ' Proc. Phys. Soc.,' March 1901. It happens that the 

 correction to the results, when calculated in this manner, is very nearly 

 equal and opposite to that already given for the variation of the 

 temperature-gradient in the flow-tube, so that if both corrections are 

 applied the results are practically unchanged. It must be remembered, 

 however, that one of these corrections is certain and obligatory, whereas 

 the other is to a great extent a matter of taste. It would really be 

 more scientific to omit the uncertain reduction to the hydrogen scale. 



The value of the difference coefficient 1-50 in formulaj 6 is calculated, 

 assuming the boiling-point of sulphur to be 444°-.'), on the scale of the con- 

 stant-pressure air-thermometer. If we took the boiling-point of sulphur 

 to be 445° -2 (as determined by Harker and Chappuis with a constant- 

 volume nitrogen-thermometer at 560 mm. initial pressure), we should 

 find c?=l-54. This would make a difference of 4 in 10,000 in the values 

 of the specific heat at 0° and 100°. But the correction from the constant- 

 volume nitrogen scale would be much smaller, so that, by a curious 

 coincidence, the final results reduced to the hydrogen scale would be 

 almost identical with those already given. 



Table II. — Variation of Sjjecijic Heat of Water in terms of a Unit at 20° C. 



' Fhil.Mag.,\'^W 



Phil. Tram., 1862. 



