102 DRS. J. A. HAUKEK AND P. CHAPPUIS ON A 



Nor is it easy to apply to the results on the nitrogen scale the correction necessary 

 to bring them to what should have been found had we been able to employ an initial 

 pressure of 1 metre instead of 528 millims. 



As has been pointed out, all our gas temperatures are referred to the constant- 

 volume scale. The connection between this and the constant-pressure scale, and the 

 corrections to be applied to each to reduce them to the absolute gas scale, have been 

 calculated by Lord KELVIN and Dr. JOULE from their experiments on the flow of 

 gases through porous plugs. Various formulae giving this correction have, however, 

 l>een proposed by KELVIN and JOULE themselves, and by others. In a recent paper 

 by ROSE-INNES* a type of formula is deduced from the same observations, which 

 applies to the results found with all the three gases experimented upon by KELVIN 

 and JOULE. RosE-lNNEst says, "To the degree of approximation to which we are 

 working, therefore, there is no thermodynamic correction needed for a constant- 

 volume gas thermometer. There may be a correction involving squares of small 

 quantities, which would appear on a nearer approximation. Such a correction, how- 

 ever, would not be worth taking into account in the case of a thermometer constructed 

 with air or hydrogen, as the unavoidable errors of experiment would certainly be much 

 larger than the correction." 



Our result for the boiling-point of sulphur is about 07 higher than that of 

 CALLENDAR and GRIFFITHS, but it may be well to point out here that the two values 

 are not necessarily inconsistent. The value of CALLENDAR and GRIFFITHS is given as 

 444'53 for the boiling-point of sulphur on the constant-pressrue air scale, the air 

 being taken under an initial pressure of 76 centims. 



Our value 445 0- 27 we give as the equivalent of the same temperature on the scale 

 of the constant-volume nitrogen thermometer, the nitrogen being taken under the 

 initial pressure of 528 millims. It is impossible, we think, at present to say from 



hydrogen thermometers varies directly as the initial pressure. Consequently, in the comparisons between 

 100" and 200, where the initial pressure was about 800 millims., the difference between the two scales 

 would be diminished to about four-fifths, and in the comparisons between 200 3 and 455 to about half of 

 what it would have been with an initial pressure of one metre. 



We may also remark that the coefficient of dilatation of air under constant pressure 



a = 0-003 674 9, 



determined by CALLENDAR and GRIFFITHS and employed by them to calculate the temperatures in their 

 observations, is sensibly higher than that which results from the experiments of REGNAULT, 



a = 0-003 670 0, 

 or the value obtained some time ago by one of us, 



a = 0-003 670 8. 



The adoption of these latter values would raise the result of CALLENDAR and GRIFFITHS about half 

 a degree.] 



* 'Phil. Mag.,' March, 1898. 

 t Jw. fit., p. 293. 



