OF ARTS AND SCIENCES. 113 



(4) Since all the above results are simple statements of the facts 

 of experiment, and are independent of formulas, we conclude that the 

 formula of Dulong and Petit (which from well-conducted experiments, 

 like those of M. Violle, deduces conclusions which trial disproves) 

 must be itself wrong. Further, since this formula contains no term 

 depending on the wave-length, it takes no account of the difference 

 here proved to exist between the relative quantities of heat and light 

 radiation from sources of high temperature, and is thus found espe- 

 cially untrustworthy at those temperatures at which it has been most 

 frequently applied. 



I do not yet venture an opinion of my own on the real temperature 

 of the sua, further than that I think it much higher than has been of 

 late believed. 



The preceding observations and inferences all seem to point to the 

 use of the highest attainable terrestrial temperatures (e. g. that of 

 the electric light) in comparisons (and the consequent least depend- 

 ance on formulas) as the safest line for future investigation. 



vol. xiv. (n. s. VI.) 



