140 Mr. J. Rose-Innes on the Practical Attainment of 



volume method; the pressure at freezing-point for the 

 hydrogen thermometer was 100 centim. of mercury, and the 

 pressure at freezing-point for the nitrogen thermometer was 

 99*5 centim. of mercury. The following table gives the 

 difference in the readings of the two thermometers (i.) as 

 observed by M. Chappuis, and (ii.) as calculated by the theory 

 given in this paper. 



Tern- 

 pr.rature. 



00 



(ii.) 



Tem- 

 perature. 



(i.) 



(ii.) 



0°C. 











60° O 



•005 



•0024 



10 



•006 



•0010 



70 



•001 



•0020 



20 



•010 



•0018 



80 



-•002 



•0015 



30 



•011 



0023 



90 



-•003 



•0008 



40 



•010 



•0025 



100 











50 



•009 



•0025 









It will be remarked that the differences observed by 

 M. Chappuis are considerably larger than the differences 

 predicted by the theory. Even making allowance for the 

 great difficulty in observing the experimental data required 

 for the calculations, it seems hard to believe that the theory 

 here put forward can be subject to such an enormous per- 

 centage error as would be implied in accepting M. Chappuis's 

 figures. It seems more probable that M. Chappuis's measure- 

 ments are subject to some unknown source of error ; and, 

 as suggested above, this may be very well due to the same 

 physical cause as that producing the difference of 0°*2 C. in 

 the estimates of the freezing-point. In order to exhibit the 

 plausibility of this notion we may have recourse to a graphical 

 method. Let us imagine that the 0°*2 C. in question is due 

 to a slight excess of pressure in the hydrogen measured at the 

 boiling-point ; not that there is any evidence to render this 

 supposition more likely than others that could be suggested, 

 but it serves to give definiteness to our construction. Then, 

 if this excess of pressure is introduced quite regularly as the 

 gas is heated from the freezing-point to the boiling-point, 

 there would be no interference with the reduced reading of 

 the thermometer : this state of things is represented by the 

 straight line in the subjoined diagram. But if this excess 

 of pressure is introduced somewhat irregularly, as represented 

 by the curved line in the subjoined diagram, we should have 

 accounted for M. Chappuis's observations. 



