WHY A FLAME EMITS LIGHT 97 



nomenon. Sunlight which was reflected and refracted by 

 luminous flames was found to exhibit phenomena identical 

 with that reflected and refracted by non-luminous flames 

 rendered luminous by the introduction of solid matter, 

 and also with light reflected and refracted by very finely 

 divided solid matter held in suspension in a liquid. The 

 phenomena presented by like experiments with glowing 

 vapors were totally different. All of Bureh's work was 

 confirmed by Stokes some years later. 



There was now left no shadow of doubt about carbon 

 being the source of the light rays, and the next question 

 that concerned investigators was the chemical changes 

 which give rise to carbon particles. 



Sir Humphry Davy thought the separation of carbon 

 to be due to a decomposition of the hydrocarbon compounds 

 (of which all illuminants are composed) within the flame 

 where the air is in smallest quantity, and no other cause 

 was assigned by other investigators. Prior to 1861 the 

 view, it seems, was that carbon is liberated because of a 

 supposed greater affinity of oxygen for the hydrogen of the 

 hydrocarbon than for the carbon, there not being enough 

 for both. But these points had to be tested. 



In the study of the chemical changes that take place, a 

 flame burning at a circular orifice offered the best condi- 

 tions. As explained in text-books of chemistry, such a 

 flame may be thought of as being made up of an inner, 

 faintly luminous cone fitting into an outer, brightly lumi- 

 nous one as a finger fits into a glove finger this latter 

 being surrounded by a non-luminous sheath of water vapor 

 and carbon dioxide. It was desirable to separate these 

 two cones, in order to study the gas after it had left the 

 inner cone and before any change had been brought about 

 by the conditions existing in the outer cone. This separa- 

 tion was first accomplished by Techlu, in France, and 

 Arthur Smithells, in England, working independently, 

 with a piece of apparatus, the essential features of which 

 are pictured in cross section in Fig. 1. By a proper con- 

 7 



