of heat, and would therefore not convey away the heat of the flame so rapidly as metal. Experiment shows, however, that the difference is very slight. It is measurable, but not of practical importance. The practical value of a lava-tip lies simply in the fact that it does not rust. Neither are any arrangements for heating the burner likely to effect any gain in illuminating power. Obviously the heat for this purpose must not be taken from the flame itself, but must either be obtained from its waste heat or from a separate source of heat. The waste heat of the flame would probably not heat the burner hot enough to produce any appreciable effect, while a separate source of heat involves additional expense enough to overbalance any gain likely to accrue.
A flame may be cooled in another way than by means of some cold body held in it, namely, by mixing some indifferent gas with the combustible gas. If, instead of air, a stream of carbonic-acid gas be admitted at the bottom of a Bunsen lamp, the flame becomes "non-luminous." We get the same amount of heat in the flame as before, since we burn the same amount of gas; but this heat has not only to be applied to heating the illuminating gas and the products of its combustion, but has also to heat the carbonic-acid gas. As a consequence, no part of the flame can be so hot as it was before, and the separation of carbon does not take place.
If the tube through which the gases pass be heated, the heat thus added counteracts the cooling effect of the carbonic-acid gas, and the flame becomes luminous again.
2. Dilution.—The second way in which we may render a gas-flame "non-luminous" is by dilution, either of the gas or the air. This can be shown by substituting for the carbonic-acid gas some combustible gas which does not give a luminous flame; for instance, carbonic-oxide gas. This gas can not essentially lower the temperature of the flame, for it burns with as hot a flame as gas, yet the mixture of it with illuminating gas gives a "non-luminous" flame. Plainly, then, we must conclude that simple dilution is able to produce this effect. Apparently such a mixture as this—that is, a mixture of two gases, one of which burns with a luminous and the other with a "non-luminous" flame—requires to be made hotter in order to produce a separation of carbon, and consequently a luminous flame, than does the illuminating gas when burned alone. If we raise the temperature of the flame by heating the burner, it becomes luminous again. Dilution, then, renders a flame "non-luminous," because the temperature of the flame is not high enough to cause a separation of carbon from the diluted gas.
3. Too Rapid Oxidation of the Carbon.—The third method of rendering a flame non-luminous is to supply it with so much oxygen that the carbon is burned at once before it can separate in the solid form.
If we put a small gas-flame into a jar of pure oxygen it is rendered
