WHY A FLAME EMITS LIGHT 95 



about nine tenths of one per cent, of hydrogen, which can 

 be separated from it only at high temperatures in an 

 atmosphere of chlorine. 



Is or did Frankland's view that glowing, dense vapors 

 cause the light appeal to Heumann, who thought it unlikely 

 that such dense vapors exist in a flame or that there is a 

 sufficiently high temperature to cause them to glow. He 

 knew, of course, that at a temperature like that of an 

 electric arc many gases do glow and give continuous spectra, 

 and that a highly heated gas under pressure acts likewise ; 

 but he argued that if carbon really does exist as such in a 

 flame, it most probably is the source of luminosity. To 

 prove its presence or absence he studied the effects upon a 

 flame of heating and cooling it, of diluting and varying 

 the temperature of the gases supplied to it, its transparency 

 and the shadows cast by it, as well as other phenomena ; and 

 the results of his experiments led him to give unqualified 

 support to the theory of Davy. 



Some account of the salient features at least of Heu- 

 mann 's elaborate investigation must be given in order to 

 convey any idea of his part in firmly fixing the " solid 

 particle" theory. By allowing a luminous flame to play 

 upon a surface which rapidly conducted heat away from 

 it, like a platinum dish, its luminosity was destroyed. Heat- 

 ing the upper surface of the dish restored the luminosity, 

 and hence Heumann concluded that cooling a flame dimin- 

 ishes its light-giving properties, while heating increases 

 them. He varied the temperature of illuminating gas be- 

 fore it reached the burner and found that the same effects 

 were produced. The heating in some cases increased the 

 normal light-giving power as much as a hundred and 

 twenty-five per cent. Further investigation showed that 

 luminosity can also be diminished or destroyed by rapid 

 oxidation of the hydrocarbons, as well as by diluting them 

 with a neutral gas like nitrogen or carbon dioxide; the 

 effect of dilution being to necessitate a higher tempera- 

 ture for luminosity. He next rendered a flame non-lumi- 



