622 



TRANSACTIONS OF SECTION A. 



instrumeut. This constant is usually called the 8 of the thermometer.^ The 

 determination of this 8 is made by taking the resistance of the thermometer at the 

 temperature of sulphur-vapour, boiling freely under the prevailing atmospheric 

 pressure. Although the value of this temperature has been determined with 

 accuracy by several different investigators for the normal pressure, its change 

 with pressure has been always a matter of considerable uncertainty. 



The authors have therefore made a redetermination of the effect of pressure 

 on the boihng-poiut, carrying the range of the experiments well outside the usual 

 atmospheric limits. Details will be given elsewhere in a later communication. 

 The results of the experiments were found to be closely represented by the 

 formula : — 



T = T, + -0904 (p - 760) - -0000519 (^j - 760)* 



where T is the temperature of the vapour on the air-scale at the pressure ^""", and 

 T, the ' normal ' boiling-poiut. 



The following table gives the value of the departure from normal boiling- 

 point at various pressures to the nearest "01° C 



It will be seen that the effect decreases somewhat rapidly with rise of 

 pressure, and is much greater than the value usually employed — viz., 0082 mm. per 

 degree, deduced from Regnault's observations. 



If, for example, the value of S be calculated for a thermometer, whose true 

 S =1-500 from observations made at 730 mm., the temperature of the vapour 

 at this pressure being obtained by use of Regnault's values for the divergence 

 from the normal temperature — the error in the 6 thus obtained would be 018. 

 This would lead to the scale of the instrumeut in use being in error by reading 

 0-36° C. high at 500° C. and 1-6° C. at 1000° C, quantities of importance for work 

 of even moderate precision. 



* This is defined by the relation 



where d is the difEerence between the ' platinum temperature ' j^t, and the ' air 

 temperature' T. 



