60 SPECTRUM ANALYSIS OF FLAMES. [MEMOIR II. 



mentary horizontal section of such a flame, it would ex- 

 hibit the appearance of a rainbow ring, and when those 

 compound rays are received on the face of a prism, the 

 constituent colors are parted out by reason of their dif- 

 ferent refrangibility, and the eye is thus made sensible 

 of their actual existence. 



When thus by the aid of a prism we analyze the light 

 coming from any portion of the superficies of a flame, we 

 in effect dissect out in a convenient manner and arrange 

 together side by side rays that have come from different 

 strata of the burning shell. These, without the prism, 

 would have pursued the same normal path, and pro- 

 duced a commixed effect as white light on the eye, but 

 with it are separated transversely, and each becomes 

 perceptible. 



It is immaterial whether we impute the light emitted 

 by an ordinary flame to the liberation of solid particles 

 of ca'bon in an ignited condition, and becoming hotter 

 and hotter as they pass outwardly towards the surface, 

 or consider these particles to be in a state of combus- 

 tion. The experiments of an ignited wire in one case, 

 and of charcoal in presence of oxygen in the other, lead 

 to the same explanation. We are not to suppose that 

 it is simply a gas which is burning; we are examining 

 the light emitted by an incandescent solid the carbon 

 particles that for the moment are set free. 



(This explanation, that the luminosity of a flame is 

 due to the temporary extrication of solid carbon, was 

 given by Sir H. Davy. It has been called in question 

 by Frankland. Experiments and criticisms have since 

 been offered by Deville, Knapp, Stein, Blockmann, and 

 others, but Davy's theory still remains substantially un- 

 affected. This is the conclusion to which Heumann has 

 come in his recently published researches on luminous 

 flames (1876).) 



