FLAMES OF COMPOUNDS OF CARBON AND HYDROGEN. 415 



It has already been stated, that certain carbohydrogen flames afford spectra 

 exhibiting bright lines separated by dark spaces. In no spectrum are these lines 

 more easily observed than in that of the Bunsen gas lamp. In order to distinguish 

 the phenomena of the spectrum which are due to different portions of the flame, it 

 is sufficient to place the lamp before a narrow vertical slit not exceeding 0*2 inch 

 in height, through which its light passes to the prism.* If the lamp be gradually 

 raised before the slit, the spectrum first seen will be derived exclusively from 

 the envelope of the flame, which reaches high above the top of the interior cone. 

 This spectrum is tolerably bright, extending without the least interruption, from 

 the line C, nearly to the line H of Fraunhofek, and exhibiting no bright line 

 whatever except the yellow line R. That line, however, is extremely flickering, 

 so as often to disappear completely ; and it seems due entirely to the yellow scin- 

 tillations which abound in the exterior envelope. When the flame is raised still 

 higher, so as to bring the top of the green cone into view, four other bright lines 

 begin to appear ; and as we continue to raise the flame so as to derive light from 

 lower and lower-portions of the flame, the bright lines become more and more 

 clearly defined, owing to the intervening spaces becoming darker ; and some fainter 

 lines become visible. At length, when light passes through the slit only from the 

 lowest portions of the flame where the exterior envelope nearly disappears, the 

 bright lines become so sharply defined as to admit of their places being ascer- 

 tained by actual measurement, with almost the same accuracy which is attainable 

 in observations of the dark lines of the solar spectrum. 



From the facility with which the lines of the carbohydrogen spectrum are 

 obtained, I conceived they might be of use in optical researches; and I soon 

 found them of great service in the prosecution of my experiments. I was, there- 

 fore, anxious to ascertain whether they belonged really to the gas flame, or were 

 caused by the accidental presence of foreign matter ; for it is well known that 

 some metals, such as copper, when present in a flame, produce bright lines in 

 its spectrum. For this purpose I burned a mixture of coal gas and air, succes- 

 sively, from an iron tube, a glass tube, a tube formed of a coil of platinum foil, 

 and the brass tube of the Bunsen lamp ; but in every case the lines remained 

 unchanged in number and position, proving that they arose entirely from the 

 combustion of the gas, and not from any matter derived from the lamp. 



On the Apparent Diversity of the Spectra of Compounds of Carbon and Hydrogen. 



Having thus studied the general phenomena of the spectra of carbohydrogen 

 flames, some of which exhibit continuous, and others interrupted spectra, we may 

 now resume the question, Whence do these differences arise ? 



It has been found by Professor Drapes that an incandescent solid body emits 



* Or we may adopt Professor Draper's ingenious mode of observing flames through a horizontal 

 slit, Phil. Mag., vol. xxxii. p. 106. 



VOL. XXI. PART III. 5 U 



