
32 MR SWAN ON THE CONSTITUTION OF FLAME. 
the difference of temperature at different points of a flame be really so great as to 
affect the quality of the light to an appreciable extent. 
This will appear, if we adopt a method of examining flame, much simpler than 
prismatic analysis, and, for the purpose, probably more delicate. I mean simply 
looking at the flame. If the doctrine of concentric rings of various colours were 
really true, we should perceive, on looking through the flame in a line passing near 
its centre, the resultant tint due to the combination of all the colours in nearly equal 
proportions. But if we looked through the edge of the flame, the line of vision would 
cut each successive coloured ring more and more obliquely as it passed towards 
the interior of the luminous matter. We should therefore look through a greater 
thickness of the inner rings than of the outer. The result would be to give a pre- 
ponderance to the inner tints, so that the apparent colour of the flame would vary 
from the middle towards the edges, at first slowly, then more rapidly, until at 
length, at the edges, we should see only the colour of the very outermostring. I 
have failed to detect any such variation of colour in the fiames I have examined. 
I therefore infer that each of the three conoids of a hydro-carbon flame has sensibly 
the same tint throughout its entire thickness. 
In my paper on the spectra of the flames of the hydro-carbons, I have described 
experiments which lead to the same conclusion, but to which I can only here refer, 
without entering into details. There is, however, one experiment, which is easily 
performed, and which I have not before described. By holding a piece of plate- 
glass over the smokeless flame of a Bunsen lamp, and looking downwards through 
the glass, a tolerably good horizontal section of the flame is obtained. It is then 
seen that both the inner blue cone and the envelope of the flame are sensibly homo- 
geneous in colour throughout their entire thickness. 
We may also verify a conclusion which I have stated in the paper already 
referred to, namely, that when common salt is placed in a flame, the envelope 
alone becomes yellow, while the inner blue cone remains unaltered in colour. 
This is strikingly apparent on looking down in the manner described, and intro- 
ducing salt into the flame. The outer mantle changes at once from purple to 
bright yellow, while the inner blue conoid remains quite unaltered in tint; and 
this is the case, whether the salt is brought into contact with the outer mantle or 
the inner conoid. Possibly this phenomenon may be explained by supposing that 
the sodium of the salt is reduced to the metallic state by the free hydrogen of 
the inner conoid, and that its vapour burns exclusively in the outer mantle, pro- 
ducing the yellow tint only in that region of the flame. 
