1897.] on the Source of Light in Flames. 371 



are two luminous effects to recognise in the combustion of phosphorus. 

 One is due to the act of formation of phosphorus pentoxide giving the 

 green glow, and the other due to the subsequent heating of the same 

 substance producing the white glow. Adopting the terminology 

 suggested by E. Wiedemann, we may say that there is chemi-lumines- 

 cence and thermo-luminescence of phosphorus pentoxide. In what is 

 ordinarily called the phosphorescence of phosphorus we have the 

 chemi-luminescence ; in the vivid combustion of phosphorus the 

 chemi-luminescence is completely overpowered and masked by 

 the thermo-luminescence. 



It is interesting to inquire how far other combustible elements 

 behave in the same way. The flame of silicon hydride may be sub- 

 jected to similar experiments. When sufficiently diluted with carbon 

 dioxide a pale greenish flame is obtained, silica being the product. 

 The green colour may therefore be attributed to the formation of 

 this compound. When the supply of carbon dioxide is reduced the 

 flame becomes brightly luminous, but the luminosity may be removed 

 by cooling with a wire ring. The optical test shows the bright light 

 to be due to solid particles, and as the glow is prolonged continuously 

 in the track of the escaping silica when a Bunsen flame is held over 

 the silicon hydride flame, it seems clear that the secondary or bright 

 luminosity of the flame is here, as in the case of phosphorus, to be 

 ascribed to a purely thermal action. The chief difference in the two 

 instances is that in the case of phosphorus hydride the product is a 

 glowing gas, and in the case of silicon hydride a glowing solid. 



Hydrocarbon flames may also be considered from the same point 

 of view, and here the facts are well known. In the first instance we 

 have to recognise in a hydrocarbon flame the bright yellow light 

 and the blue or lilac light. The bright yellow light may be 

 suppressed by cooling by means of a wire or by diluting the gas 

 with carbon dioxide. This part of the light of a hydrocarbon 

 flame has frequently been ascribed to a preferential burning of the 

 hydrogen, whereby carbon is separated in the flame and glows in 

 the state of solid particles. This view, which appears to have 

 originated in a misinterpretation of Davy's words, has never been 

 based on substantial evidence, and it is at variance with the most 

 cogent experiments on the subject. There seems little doubt that 

 the bright glow of a hydrocarbon flame is essentially a thermal 

 phenomenon. 



The glowing substance was supposed by Davy to be solid 

 particles of carbon, by Frankland to be the vapour of dense hydro- 

 carbons. These two rival views have been subject to considerable 

 discussion, especially by Heumann.* 



It seems extremely difficult now to find any good evidence for 

 the dense hydrocarbon theory. One of the simplest arguments 

 against it was supplied by Stein, who pointed out that the glowing 



* Phil. Mag. [i] 89, p. 366 (1877). 



