﻿112 M. H. St. -Claire Deville on the Temperature of Flames, 



fact in the mere increase of density which necessarily accom- 

 panies the compression of the gas. He also draws conclusions 

 which seem to invalidate the classical ideas introduced into 

 science by Sir Humphry Davy, and which withdraw from the 

 theory of flame a basis which has always appeared beyond the 

 reach of attack. I confess that on this latter point I do not 

 share Professor Frankland's ideas ; and I base my opinion on 

 certain facts as yet imperfectly investigated, but which I shall 

 describe before long, when I shall have given them that demon- 

 strative form which they want, and which in the present state 

 of science must be given to all our speculations. 



I shall not attack with the same firmness the questions relative 

 to the influence of density on the luminous power of flames. I 

 propose here to develope an idea the germ of which I find in the 

 last paragraphs of Professor Frankland's communication. Our 

 colleague ascribes the want of illuminating power in the flame of 

 phosphorus burning in chlorine to the slight elevation of tem- 

 perature which a combustion accompanied by so small a disen- 

 gagement of heat must obviously produce. I believe this to be 

 the real and only reason. 



Let us first inquire what is the principal condition for lumino- 

 sity in a flame. If we take an obscure but hot flame like that of 

 a Bunsen's burner, and introduce common salt into it, every 

 one knows that we obtain a light of feeble intensity and which is 

 monochromatic; for the prism does not decompose it into a spec- 

 trum, and only produces one bright band. But if we increase 

 the temperature of this' flame (by adding oxygen, for instance), 

 the lustre immediately revives, the number of lines increases, and 

 hence approximates to a complete spectrum. The experiments 

 of M. Fizeau and of MM. Wolf and Diacon are remarkably de- 

 finite from this point of view. But suppose we use M. Debray's 

 apparatus, by which, for spectroscopic experiments, a very high 

 temperature (one of about 2500°) can be obtained. In this 

 flame the spectrum of sodium spreads out and becomes com- 

 plete; it may be assumed, then, that the great number of bril- 

 liant lines which the spectrum contains merge into each other 

 to form a whole which seems continuous. An observation of 

 the same kind is made when large quantities of sodium are burnt 

 in air or oxygen, or when lithium is set on fire : the flame of 

 sodium, which is ordinarily yellow and monochromatic, that of 

 lithium, which is usually red, both become white ; they then 

 contain all the rays or, we may say, all the brilliant lines of every 

 refrangibility. They thus become luminous when the metal 

 burns at a high temperature. 



This observation is exact also for the invisible rays — for the 

 chemical rays whose lines crowd and multiply in the spectrum in 



