264 Proceedings of Royal Society of Edinburgh. [sess. 
When a glass jar was inverted over burning ether, the blue part, 
which does not show scattered light, extended higher till, just 
before the dame went out, the luminous part disappeared altogether. 
A Bunsen flame, fed with chloride of sodium, did not show the 
phenomenon, though the flame was fairly luminous. 
The phenomenon shows very prettily the separation of carbon 
(associated, it may be, with some hydrogen) in the flame, and at 
the same time the extreme thinness of the layer which this forms. 
It shows, too, the mode of separation of the carbon, namely, that it 
is due to the action of heat on the volatile hydrocarbon or vapour of 
ether, as the case may be. At the base, where there is a plentiful 
supply of oxygen, the molecules are burned at once. Higher up the 
heated products of combustion have time to decompose the com- 
bustible vapour before it gets oxygen enough to burn it. In the 
ether just going out, for want of fresh air, the previous decomposi- 
tion does not take place, probably because the heat arising from the 
combustion is divided between a large quantity of inert gas (nitrogen 
and products of combustion) and the combustible vapour, so that the 
portion which goes to the latter is not sufficient to decompose it 
prior to combustion. 
In the Bunsen flame fed with chloride of sodium, the absence of 
scattered light tallies with the testimony of the prism, that the 
sodium is in the state of vapour, though I would not insist on this 
proof, as it is possible that the test of scattering sunlight is not 
sufficiently delicate to show the presence of so small a quantity of 
matter in a solid or liquid state. — Yours, sincerely, 
G. G. Stokes. 
P.S — I fancy the thinness of the stratum of glowing carbon is 
due to its being attacked on both sides — on the outside by oxygen, 
on the inside by carbonic acid, which with the glowing carbon 
would form carbonic oxide. 
