100 DBS. J A. HARKER AND P. CHAPPUIS ON A 



variation of pt with pressure, from which, by combination with the known value of 

 d.pt/dt, we obtained two concordant values for dt/dp at 760 millims. The mean of 

 these values coincided sufficiently nearly with that of RE<;XAM,T to justify us in 

 adopting the latter for present purposes, and the reduction of our observations to 

 normal pressure is therefore based on the assumption of his value.* 



We found that for the thermometers K.8 and K.9 the value of d.pt/dt, at the 

 sulphur point, was practically identical with the mean of those previously obtained 

 for the older thermometers, and by assuming this number and combining it with the 

 value calculated from REGNAULT'S experiments for dt/dp, we obtained d.pt/dp tor 

 K.8 and K.9. 



From this the value of pt,, the platinum temperature of the sulphur vapour at 

 760 millims., was then calculated for each experiment, and the mean value for each 

 series taken. 



The platimnn temperatures found in these comparisons made with each thermo- 

 meter near the point 445, and their corresponding values on the nitrogen scale, were 

 then treated by least squares, to obtain from them the nitrogen temperature equivalent 

 to the value of pt, found previously. Various formulae for this calculation were tried, 

 the most suitable one being found to be 



T. = x + y (pt - pt,) + z (pt - pt.Y 



where T, is the nitrogen temperature sought corresponding to pt,, and x, y and 2 are 

 constants, t 



In this calculation for thermometer K.8 were included the seven experiments 

 numbered 85 to 91 in the table, and in the calculation weights were assigned to the 

 individual experiments according to the constancy of the temperature. On 

 substituting in the original equations of condition the greatest residual was found to 

 be 0'034, showing a satisfactory concordance between the values for T, given by the 

 different comparisons. 



* Although we do not wish to give the formula we calculated from the observations made at Kew as 

 the outcome of a new determination of dt/dp for sulphur, yet it may be worth while to give an idea of the 

 kind of agreement between the value found and that of REGNAULT, which we adopted for the reduction of 

 our observations. The experiments with thermometer K.I were made between the extremes of pressure 

 747 and 773 millims., but the majority of them were only very slightly removed from 760 millims. The 

 series with K.3 was better adapted for the purpose of deducing a formula, the observations being 

 distributed fairly evenly over the range 747 to 769 millims. These two sets of experiments were made by 

 Mr. HUGO, Senior Assistant at Kew Observatory. 



If T g be the boiling-point under 760 millims. pressure, we have for the value at 755 millims. from the 

 formulae deduced from REGNAULT, and from thermometers K.I and K.3, the values (T,- -41), (T,- -43), 

 and (T, - "42), respectively. 



t For another method of arriving at the value of pt, and the corresponding T, leading to a mean result 

 slightly different from that here given, see Appendix II., added while the paper was in press. 



