Mr. B. Stewart on the History of Spectrum Analysis. 355 



so that here we are at once led to acknowledge a radiation pro- \ 

 ceeding from the interior of bodies as well as one from their ) 

 surfaces. 



Having thus shown that the equilibrium demanded by Pre- 

 vost's theory extends into the substance of bodies, experiments 

 were next made on rock-salt, mica, and glass, which resulted in 

 proving " that every body which sifts heat in its passage through 

 its substance is more opake with regard to heat radiated by a 

 thin slice of its own substance, than it is with regard to ordinary 

 heat. - " This experimental result was conceived to involve the fol- 

 lowing extension of the law of equilibrium : " The absorption of 

 a plate or particle equals its radiation, and that for every descrip- 

 tion of heat. v 



On this proof from experiment Kirchhoff remarks as follows : 

 " The proof cannot be a strict one, because experiments which 

 have only taught us concerning more and less, cannot strictly 

 teach us concerning equality." 



I presume this means that since one cannot experimentally 

 insulate a ray of heat of a definite refrangibility, a strict proof 

 cannot in this w r ay be obtained. I shall only remark that the 

 same objection holds with regard to many of our most valuable 

 experiments on dark heat. I cannot, however, admit that the 

 difficulty of this experimental verification is surmounted or even 

 lessened by employing a high temperature and selecting a defi- 

 nite ray of light. For in this case it is necessary to compare 

 together at the same temperature a body such as a flame, and a 

 black body such as carbon ; and here while an equality of radia- 

 tion and absorption may be easily procured, it will be very diffi- 

 cult, if not impossible, to prove the necessary equality of tempera- 

 ture between the two bodies. I shall afterwards show that the 

 best experimental proof of this law is given by the fact that 

 coloured glasses lose their colour in the fire. 



To return to the proof from dark heat. This is followed by a 

 theoretical demonstration that the law thus rendered experiment- 

 ally probable is also necessarily true. Kirchhoff remarks regard- 

 ing this : — " And then he (Stewart) proceeds to more abstruse 

 considerations which are intended to give such a rigid demon- 

 stration, and in which the meaning he attaches to the expressions 

 absolution and radiation are more nearly defined. These consi- 

 derations, however, are not sufficiently general or sufficiently 

 precise to attain the required end ; so that, after all, Stewart^ 

 proposition remains an hypothesis to which some probability is 

 attached." 



Now this proof is so far from being abstruse, that, with your 

 permission, I shall state it simply in a few words, leaving men 

 of science to judge for themselves of its completeness. Having 



