AND ON THE COLOURS EXHIBITED BY CERTAIN FLAMES. 455 



results are of course influenced by the peculiar character of 

 this light, as those rays which the sun does not emit, or emits 

 but sparingly (as in the confines of the red and yellow), can- 

 not be found in the transmitted beam, however well disposed 

 the medium may be to give them passage. These rays, how- 

 ever, may exist in light from other sources, as star-light, elec- 

 trical or phosphorescent light, or that of flames, which differ 

 extremely in their types when examined with the prism, and 

 that in a manner the most capricious imaginable. I shall here 

 set down a iew examples of remarkable flames which I have 

 observed. 



Sulphur, when burning with its usual feeble flame, emits all 

 the rays, but principally the violet and blue. When the in- 

 flammation is violent, however (as when a piece of sulphur is 

 thrown into a white-hot crucible), the light emitted is perfect- 

 ly homogeneous, being a pure, brilliant yellow, and of a re- 

 frangibility so strictly definite, that all the minute flickerings 

 of its flame, seen through a prism of the largest refracting 

 angle, appear as sharply defined, and free from nebulosity, as 

 when viewed with the naked eye. As the violence of the in- 

 flammation abates, a faint train of green and momentary red 

 spectra appears. 



To insure a regular and violent inflammation of the sulphur, 

 I attempted to burn it with nitre, but the definite character of 

 the flame was now lost, other colours coming in, and, in parti- 

 cular, two red spectra, sharply separated from the yellow and 

 each other, as in Fig. 11., in which the dotted curves repre- 

 sent the additional portions of the type arising from the nitre. 



14. The flame of alcohol (a common spirit-lamp), consists 

 of two portions, — a cone of yellow flame, inclosed in a shell or 

 envelope of blue, but projecting above it, like an acorn from 



vol. ix. p. ii. 3 m its 



