258 THE POPULAR SCIENCE MONTHLY. 



the spectral bands of the electric arc (which are identical with 

 those of the candle and the oxyhydrogen blow-pipe) may be car- 

 ried, I have published a specimen of the fades of the yellow, 

 green, and blue bands, indicating the intensity and normal dis- 

 tance of the rays composing them. It appears, from this work, 

 that, for a fifth part alone of their total length, these bands show, 

 respectively, 163, 160, and 120 lines ; this would bring to about 

 800 the number of lines constituting each band, and to at least 

 4,000 the number of the lines forming the five bands of the spec- 

 trum of the electric arc ; for the more intense bright lines are 

 doubled again when they are observed under conditions favor- 

 able to their brilliancy and dispersion. In comparing, with the 

 same spectroscope, the spectrum of the electric arc and the solar 

 spectrum, we observe that the former spectrum displays a more 

 considerable number of bright rays than the solar spectrum of 

 dark rays. Since it is nearly certain that the spectral bands be- 

 long to the spectrum of carbon — for they are observed when the 

 electric arc shines in a vacuum, that is, when carbon alone is in 

 ignition — it follows that the spectrum of this element contains 

 more rays than the entire solar spectrum. 



Some physicists doubted for a long time the identity of the 

 spectra of carbon and the candle-flame, because there existed a 

 spectrum of carbon entirely different from the banded spectrum. 

 But as I have succeeded in demonstrating, on the one hand, that 

 this spectrum does not belong to carbon, and on the other hand 

 that the spectrum of the candle-flame was brightly visible in the 

 ignited filament of the incandescent lamp when the vacuum is 

 as perfect as it is possible to make it, I think there should now 

 be little doubt respecting the identity of the two spectra. Car- 

 bon, being found in various combinations everywhere on the sur- 

 face of the globe, should of necessity reveal its presence in most 

 of the bodies subjected to spectrum analysis. Eminent chemists 

 have even found traces of it in the nearly perfect vacuum of our 

 pneumatic machines. 



The absorption spectrum of carbon, or that which should be 

 composed of the dark lines detaching themselves upon a contin- 

 uously bright spectrum, has not yet been obtained. In the com- 

 parative study that I have made of the solar spectrum and the 

 spectrum of carbon, I have shown that most of the bright rays 

 forming the carbon bands do not coincide with the dark rays of 

 the solar spectrum. I have been inclined to believe from this 

 that the absorption spectrum of carbon does not exist in the so- 

 lar spectrum, but I have not been able to declare the same con- 

 clusion respecting the emission spectrum — that is, the spectrum 

 with bright bands — because the discovery of the bright bands in 

 the solar spectrum offers a real difficulty, resulting from the fact 



