110 



Mr. E. F. J. Love on the 



the following values were deduced, 



7 = 4-11 xlO 5 ; 8 = 0-88. 



With these constants the values of 6 were recalculated for 

 the experimental temperatures, and the annexed table gives 

 the results. It will be seen that, slight as are the differences 

 between the results of observation and those furnished by 

 Rose-Innes's formula, the differences in the case of the new 

 formula are still smaller, as would naturally be anticipated 

 from the above discussion. 



Table. 



Temp. 

 Centigrade. 



o 



00 



8 (observed). 



(calculated) 

 (Rose-Innes). 



Differences 

 (obs. — calc). 



9 (calculated) 

 (new formula). 



Differences 

 (obs. — calc). 



461 



4-o0 



+ 004 



4-64 



00 



74 



437 



435 



+0-02 



435 



+0-02 



35 6 



3-41 



349 



-0-03 



343 



-002 



51-0 



2-95 



3-02 



-0-07 



296 



-001 



93-5 



2-16 



216 



00 



218 



-002 



976 



214 



2-08 



+0-06 



212 



+0-02 



Several remarks are suggested by the discussion just given. 



In the first place, we observe that the Joule-Kelvin formula 



for the thermal effect would hold good for gases whose 



co-volume was zero, or negligible in comparison with the 



" molecular pressure " ; it is not unlikely, indeed, that if the 



1 



original investigators had plotted 6 and y^ 2 , instead of ^ 



and T, they would have arrived at the formula here suggested. 

 Be that as it may, their suggestion that the temperature 

 enters into the expression in its second power is probably very 

 near the truth. 



Another point worth noticing is the testimony, borne by 

 the calculations, to the remarkable accuracy of these experi- 

 ments ; this statement holds good whether Rose-Innes's 

 formula or the present writer's be adopted, though most 

 markedly, perhaps, in the latter case. This remark seems 

 worth making as, in the present writer's opinion, too little 

 credit has in the past been attached to the experiments in 

 this particular ; for example, Lehfeldt's statement*, that we 



* Phil. Mag. April 1898. 



