THE FLAME OF A CANDLE. 257 



Tlie briglit envelope nearly all disappears^, while the inner dark 

 cone is considerably developed, reaches a very high temperature, 

 and exhibits a spectrum identical with that of the outer cone of 

 the original flame. The brightness of the spectral bands is aug- 

 mented by the rising of the temperature, and two new luminous 

 bands, a red and a violet one, become visible in the spectro- 

 scope. At this moment we recognize that these bands are com- 

 posed of a series of rays or bright lines, separated from one 

 another by dark spaces. 



While performing the prismatic analysis of the inner cone 

 of a gas-light flame fed with pure oxygen, M. Stas observed, with 

 the same spectroscope, a spectrum sensibly different as to the 

 number of rays constituting the bands, according as the obser- 

 vation was made upon the top of the inner cone, where the tem- 

 perature is highest and sufficient to keep iridium in fusion, or 

 on the front or the side of this inner cone. The physiognomies 

 of these three spectra vary according to the spectroscope em- 

 ployed. If we use a spectroscope with direct vision and weak 

 dispersion, we observe a spectrum resembling that of the candle- 

 flame ; but, -with an instrument of more considerable dispersive 

 power, the bands define themselves into brilliant rays, some fine, 

 and others broad, having extremely clear edges. These facts, M. 

 Stas remarks, inseparably connect the fades of the spectrum of 

 the flame with its greater or less elevation of temperature, and 

 with the analyzing instruments employed. 



Although the luminous intensity of the inner cone of the 

 oxyhydrogen-flame is quite weak, Mr. Piazzi Smyth has discov- 

 ered more than 400 bright rays in the spectral bands of this cone ; 

 viz., 97 rays in the red, 94 in the yellow, 97 in the green, 107 in 

 the blue, and 71 in the violet bands. 



But it is the analysis of the electric arc, the light of which 

 does not differ essentially from that of the candle — for it is also 

 the result of the ignition of carbon — that shows us these spectral 

 bands in all their splendor, and initiates us into the grand com- 

 plexity of their constitution. Like a luminous ribbon passing 

 insensibly from one shade to another with diminishing brill- 

 iancy, each band is composed of a considerable number of bright 

 rays of different breadths, disposed with a wonderful symmetry, 

 increasing with the power of the analyzing instrument and the 

 luminous intensity of the electric arc ; the broader bright rays 

 doubling into finer rays, and new luminous rays appearing in 

 the dark spaces that separate the bright rays. While these 

 bright lines are not arranged rigorously in the same manner in 

 each band, they nevertheless show a great resemblance in their 

 grouping and spacing. 



In order to show how far the resolution into bright lines of 



VOL. XXXIII. — 17 



