254 Dr. W. M. Watts on the Spectra of Carbon, 



trace of a, 7, 8, e. Fig. 2, c and d, shows the result of a direct 

 comparison of the carbonic-oxide vacuum-spectrum with that of 

 the olefiant-gas flame when the two are seen simultaneously in 

 the spectroscope. 



The carbonic-oxide vacuum-spectrum shows the lines h,j, k, 

 /, m, ?i, and o. A spectrum-tube enclosing olefiant gas (or coal- 

 gas, or a mixture of equal volumes of olefiant gas and hydrogen) 

 gives h, j, k, I, m, n, and o, and sometimes the group 6 of the 

 first carbon-spectrum ; occasionally 8 is also faintly visible. 

 Plucker* has obtained from a vacuum-tube containing carbonic 

 disulpkide, carbon h,j, k, /, m, and n. 



I believe that we have a third form of the carbon-spectrum in 

 that obtained from the Bessemer-flame, which I described in a 

 paper published in this Magazine for December 1867. Professor 

 Lieleggf regards the Bessemer- spectrum as that of carbonic 

 oxide. It is, however,- impossible to obtain it either from the 

 flame of carbonic oxide or from the gas rendered incandescent by 

 electricity : in the first case a continuous spectrum only is ob- 

 tained ; and in the latter either the spectrum of carbon (fig. 1) or 

 that obtained also from carbonic anhydride (fig. 4) is produced. 

 I have always looked upon this spectrum as that of carbon 

 itself, and have sought to obtain it from compounds of carbon 

 with nitrogen or with hydrogen, but without success. It appears 

 to be produced only under conditions very nearly the same as 

 those of the Bessemer-flame itself. Thus I have observed it in 

 one or two furnace-flames in which a very high temperature is pro- 

 duced. The flame of carbonic oxide in an ordinary melting-cupola 

 gives a very brilliant continuous spectrum, but exhibits only the 

 sodium-line. In the working of a blast-furnace it is usual, after 

 the iron has been run, to turn on the blast so as. to help the iron 

 out. This produces a large white flame from under the tymp, 

 which exhibits a very bright continuous spectrum with the so- 

 dium- and lithium-lines brilliant, together with a faint Bessemer- 

 spectrum. I have observed the lines of the Bessemer-spectrum 

 also in the flame of a small furnace, used on the works at Crewe 

 for loosening the tyres of wheels, in which coke is burnt by a 

 blast of air ; and the Bessemer-spectrum is always obtained in 

 the combustion of coke alone in the convertor. The spectrum 

 of the coke-flame exhibits the Bessemer-lines faintly, and the 

 lines of sodium and lithium : the introduction of the charge of 

 molten pig iron seems to cool down the flame, so that for two or 

 three minutes after the commencement of the blow a continuous 

 spectrum only is seen. As the temperature rises the sodium- 



* Pogg. Ann. vol. cvii. p. 538. 



f Phil. Mag. S. 4. ypl. sxxiv. p. 302. 



