58 SPECTRUM ANALYSIS OF FLAMES. [MEMOIR II. 



introduced, two totally different things are confounded, 

 and that the burning coal not only gives forth its rays 

 as a combustible body, strictly speaking, but also as an 

 incandescent mass. 



To avoid this objection as far as possible, and also to 

 reach a much higher temperature than could have been 

 otherwise obtained, I threw a stream of oxygen gas on 

 that part of the anthracite which was opposite the slit ; 

 but my expectations were disappointed, for, instead of 

 the combustion being increased, the coal was actually 

 extinguished by the jet playing on it. I therefore re- 

 placed the anthracite with a flat piece of well-burned 

 charcoal, kindled at the portion opposite the slit, and 

 throwing a stream of oxygen on this part, the combus- 

 tion was greatly increased. A spectrum rivalling that 

 of the sunbeam in brilliancy was produced, all the col- 

 ors from the extreme red to the extreme violet being 

 present. 



On shutting off the supply of oxygen, the combustion, 

 of course, declined, and while this was going on the vio- 

 let, the indigo, the blue, the green, etc., faded away in 

 succession. By merely turning the gas on or off, the 

 original colors could be re-established or made to decline. 

 It was very interesting to see with what regularity, as 

 the chemical action became more intense, the more re- 

 frangible colors were developed, and how, as it declined, 

 they disappeared in due succession; the final tint being 

 red and that ash-gray in the position of the yellow which 

 has been described in the preceding memoir. 



In the form of experiment here made the combustion 

 is, of course, merely superficial ; and the rays come from 

 the charcoal not as an incandescent, but as a burning 

 body. 



III. Of tlu constitution of flames; proving that they 



