( 411 ) 



XXIX.— On the Prismatic Spectra of the Flames of Compounds of Carbon and 

 Hydrogen. By William Swan, F.R.S.E. 



(Read 21st April 1856.) 



The phenomena presented by the prismatic spectra of flames have occupied 

 the attention of many and excellent investigators. In most instances, however, 

 no attempt has been made to procure accurate measurements of the positions 

 of the bright lines which many of the spectra exhibit ; and much in this field 

 of observation, therefore, remains to be accomplished. I purpose, from time to 

 time as I shall have leisure, to make a series of observations, whose object shall be 

 the actual numerical determination of the positions of the bright lines in the 

 spectra of flames ; and I have commenced the series with an examination of 

 the spectra of the flames of compounds of carbon and hydrogen. In an in- 

 vestigation into the phenomena of flames, the compounds of carbon and hydro- 

 gen claim our first attention, as constituting the most important means of 

 artificial illumination; for it is scarcely necessary to remark, that, with the 

 grand exception of sun-light, the combustion of these substances is the source 

 of nearly all the light and heat from which we derive such extensive benefits 

 in the arts and in domestic economy. It will be found, moreover, that the spectra 

 of carbohydrogen flames possess, in common, remarkable features, which seem as 

 yet to have received little attention, but which promise to be of service in ex- 

 plaining the general phenomena of artificial light. 



If we examine the spectrum of the brightest part of the flame of an oil-lamp 

 or a tallow-candle, it will be found that it exhibits no dark intervals, and that 

 its colour and brightness vary gradually from point to point with scarcely any 

 breach of continuity. If, however, we observe only the light proceeding from the 

 blue part of the flame, which surrounds the upper part of the wick, a totally 

 different result is obtained. The extreme red and violet rays become nearly or 

 altogether invisible, and the intermediate portion of the spectrum exhibits a 

 series of bright lines separated by dark intervals.* Similar lines occur in 

 the flames of alcohol, sulphuric ether, and wood spirit. They are seen, how- 

 ever, with great difficulty in the flame of impure wood spirit, and are scarcely, if 

 at all visible, in the more luminous flames of oil of turpentine and coal naphtha. 

 Before offering an explanation of these differences, it will be necessary to premise 

 some particulars regarding the nature of flame. 



* For descriptions of lines in the spectrum of an oil lamp flame, see Fraunhofer, Astronomische 

 Abhandlungen, 1823, p. 16; Herschel, Edin. Trans., vol. ix., p. 455, and article Light, Encyc. 

 Metrop., art. 522. The spectrum of the blowpipe cone is described by Fraunhofer, Brewster's 

 Journal of Science, vol. vii., p. 7 ; and by Draper, Phil. Mag., vol. xxxii., p. 111. 



VOL. XXI. PART III. 5 T 



