670 PHILOSOPHICAL TRANSACTIONS. [ANNO 1785. 



the green, in the greatest abundance. When the combustion of the spirits was 

 checked by throwing some sal ammoniac into the mixture, the red rays disap- 

 peared ; but when, by the long continuance of the flame, the sal ammoniac was 

 rendered so hot as to increase, rather than diminish the combustion, the red rays 

 again appeared at the perforation. If the screen was managed so that the differ- 

 ent parts of the flame might be examined separately, I always observed that the 

 colours varied according to the degree of heat. At the base of the flame, or 

 where the heat was least, the indigo, the violet, and a very small tinge of the 

 blue and green appeared. As I approached the vertex of the flame, the rays 

 which escaped became more and more numerous, till I reached the top, when all 

 the rays appeared in the prism. When the red rays first made their appearance, 

 their quantity was small, and gradually increased as the eye in its examination 

 approached that part where the heat was greatest. Mr. Melvill, when he made 

 some of the preceding experiments, observed, that the yellow rays frequently- 

 escaped in the greatest abundance ; but this singularity proceeded from some cir- 

 cumstances which escaped his attention. In consequence of mixing acids or salts 

 with the burning spirits, a very dense fume of unignited particles arises, and be- 

 fore the rays of the burning body arrive at the perforation where the prism catches 

 them, they must pass through a medium which will absorb a great part of the in- 

 digo and the violet. On the other hand, owing to the imperfection of the de- 

 composition, very few of the red rays are separated from their basis, and con- 

 sequently the yellow and the orange rays are those alone which pass through the 

 unburnt smoke of the flame. 



But, besides the increase or decrease of heat, there are other modes of retard- 

 ing or accelerating the combustion of bodies, by which also may be examined 

 some of the preceding illustrations. 1. A candle burns most rapidly and bril- 

 liantly in dephlogisticated air. 2. The blue colour of a sulphureous flame in 

 pure air is changed into a dazzling white. 3. The flame of inflammable air 

 when mixed with nitrous air, is green. It is white strongly tinged with the in- 

 digo and violet when mixed with common air ; but when mixed with dephlogisti- 

 cated air, or surrounded by it, the brilliancy of its flame is most singularly 

 beautiful. 



If the preceding facts prove that light, as an heterogeneous body, is gradually 

 decomposed during combustion ; if they prove likewise, that the indigo rays 

 escape with the least heat, and the red with the greatest; I think we may 

 rationally account for several singularities in the colours of different flames. If 

 a piece of paper, impregnated with a solution of copper in the nitrous acid, be 

 set on fire, the bottom and sides of the flame are always tinged with green. 

 Now this flame is evidently in that weak state of decomposition, in which the 



