MEMOIR II.] SPECTKUM ANALYSIS OF FLAMES. 9 



introduction of a jet of air by a blow-pipe should make 

 the combustion rapid where before it was slowest, and 

 the less refrangible colors ought to be destroyed. A 

 prismatic analysis should exhibit the spectrum of a 

 blow-pipe flame without any red or orange. 



In this examination no slit was required, as in the 

 former experiments, for the cone itself, when at a dis- 

 tance of six or eight feet, was narrow enough for the 

 purpose. It yielded a very extraordinary spectrum. As 

 I anticipated, all the red rays were gone; not a vestige 

 of either them or of the orange could be found. But 

 the spectrum was divided into five well-marked regions, 

 separated from one another by dark spaces. There were 

 five distinct images of the blue cone : one yellow, two 

 green, one blue, and one violet. In Fig. 9 this result is 

 represented. 



This experiment may be verified without a telescope. 

 On looking through a prism, set horizontally at its angle 

 of minimum deviation, at a blow-pipe cone some six or 

 eight feet distant, there will be seen a spectrum of that 

 part of the flame which does not join in the production 

 of the blue cone. It contains, of course, all the prismatic 

 colors. But projecting from this are five colored images 

 of the cone one yellow, two green, one blue, and one 

 violet. They are entirely distinct from one another, and 

 are parted by dark spaces. 



Such is the effect of introducing air into the interior 

 of a flame, and destroying those 

 strata that yield the red and 

 orange colors. The effect of a 

 blow -pipe is to produce two 

 strata of blue light, one being 

 external, the other internal; also F . y 



two strata of green, one again 

 external, the other internal, and the escaping products 



