FLAMES OF COMPOUNDS OF CARBON AND HYDROGEN. 419 



Comparison of the Carbohydrogen and Solar Spectra. 



Having ascertained, that probably all substances of the forms C r H s and C r H B O t 

 produce, when burning, spectra which are absolutely identical, I was desirous to 

 compare their spectra with that of sun light. 



For this purpose I at first attempted to view the solar spectrum and that of 

 the Bunsen lamp simultaneously, but the great comparative faintness of the lat- 

 ter rendered that mode of comparison exceedingly difficult. I therefore deter- 

 mined to measure separately the minimum deviations for the principal lines of 

 the solar and gas spectra; the intervals between the adjacent smaller lines of the 

 latter spectrum being ascertained by means of a micrometer. 



The instruments I employed were an excellent theodolite by Adie, and a very 

 fine flint-glass prism by Secretan of Paris, whose faces have an area of four 

 square inches, and which shows, with great distinctness, the finest lines in Fraun- 

 hofer's map of the solar spectrum. I am indebted for the use of these instru- 

 ments to the kindness of Mr John Adie and Professor Forbes. 



The prism being placed in its position of minimum deviation, the indices of 

 refraction given in the sequel were calculated by the formula 



._ sinHl + D) . 

 P ~ sin J I ' 



where I is the angle of the prism, and D the deviation of the transmitted light. 



I have denoted the five brightest lines of the carbohydrogen spectrum by the 

 letters, a, 3, 7, 8, £ ; and the fainter lines by which they are accompanied by (3 V /3 2 , 

 7 15 &c. In the tables, D A , D y , fi A , /j. y , &c, denote respectively the minimum devia- 

 tion of the rays, and the index of refraction for the lines A and 7 of the solar 

 and flame spectra. 



A comparative diagram of the spectra of sunlight and the hydrocarbon flames 

 is given in Plate VIIL, fig. 1, where a is the double yellow line R of Fraunhofer. 

 I have thought it advisable to introduce this line in the diagram, as it is almost 

 constantly visible in ordinary artificial light, although, for reasons already fully 

 stated, I conceive it is not peculiar to the spectra of carbohydrogen compounds. 

 This conclusion is strongly corroborated by the remarkable phenomena pointed 

 out at p. 414, namely, that the salts of sodium tinge the exterior envelope of 

 the Bunsen lamp flame with so brilliant a yellow light, as completely to over- 

 power the comparatively feeble blue light of the inner cone, and to render it al- 

 together invisible ; while yet the light of that portion of the flame remains abso- 

 lutely unchanged. This remarkable property of flame is easily demonstrated by 

 holding a slip of platinum, with some salt placed on it, in the flame, while the 

 spectrum is observed through a telescope. The instant the salt reaches the flame, 

 the yellow line R or a, which before may have been extremely faint, or altogether 

 vol. xxi. part in. * 5 x 



