MEMOIR II.] SPECTRUM ANALYSIS OF FLAMES. 57 



and became of a dazzling brilliancy; and on examina- 

 tion through the telescope, though all the colors ha'd 

 increased in brightness, the most remarkable effect took 

 place among the extreme refrangible rays. Far out of 

 the limits of the ordinary spectrum a ray of great purity 

 and force was developed, as represented in the Fig. at 

 No. 5. Its color is violet. 



I have made similar experiments on many other flames 

 besides those here mentioned. It is not necessary to re- 

 late them in detail, for they give the same results. In 

 every instance of combustion in the air, when the flame 

 is bright enough, all the colors are visible; and when 

 the combustion takes place in oxygen they are increased 

 in intensity. With hydrogen gas and alcohol the light 

 is so feeble that the eye cannot catch the terminal rays ; 

 but as soon as the combustion is made in oxygen the 

 red and the violet both appear, the latter, however, pre- 

 dominating. Several of these spectra both in air and 

 oxygen are represented in Fig. 8. In No. 9, the letters 

 m g and m b indicate a maximum of green and blue 

 light in the form of bright lines. 



It does not require the use of a prism to satisfy one's 

 self of the change of tint that flames exhibit when the 

 chemical action increases. In reality it is only necessary 

 to compare by eye-sight the color of the light emitted 

 in air and in oxygen gas. In the latter case rays of a 

 higher refrangibility uniformly arise. 



On the evidence furnished by the foregoing experi- 

 ments I regard a common flame as consisting of a shell 

 of ignited matter in which combustion is going on with 

 different degrees of rapidity at different depths, being 

 most rapid at the exterior, where there is a more perfect 

 contact with the atmosphere, and diminishing inward. 

 In a horizontal section, the interior space consisting of 

 unburned vapor is black; this is surrounded by a ring 



