MEMOIR II.] SPECTRUM ANALYSIS OF FLAMES. 



55 



Fig. T. 



If, instead of a horizontal, we wish to examine a vertical 

 element of the flame, the slit and the prism must, of 

 course, be set vertically. The former mode possesses 

 great advantages, as will be presently pointed out. It 

 is to be understood in all cases that the eye-piece of the 

 telescope is adjusted to give a sharp image of the slit, 

 and the prism is at its angle of minimum deviation. 



By this arrangement I have examined a great number 

 of different flames, as those of oil, alcohol, solutions of 

 boracic acid arid nitrate of strontian in alcohol, phos- 

 phorus, sulphur, carbonic oxide, hydrogen, cyanogen, 

 arseniuretted hydrogen, etc. Among these, it will be 

 noticed that different colors occur. Oil gives a yellow 

 flame, alcohol a pale blue, boracic acid green, strontian 

 red, phosphorus yellowish - white, sulphur and carbonic 

 oxide blue, hydrogen pale yellow, cyanogen lilac, arseni- 

 uretted hydrogen white, etc. 



Notwithstanding this diversity of color, all these 

 flames, as well as many others I have tried, yield the 

 same result: every prismatic color is found in them. 

 Even in those cases where the flame is very faint, as in 

 alcohol and hydrogen gas, not only may red, yellow, 

 green, blue, and violet light be traced, but even bright 

 Fraunhoferian lines of different colors. 



