consists practically of compounds of carbon and hydrogen. The gas must get very hot before it actually begins to burn, and it is a well known fact that at a high temperature some of the compounds contained in it are decomposed, with separation of carbon.
In the interior of the flame, likewise, the compound of carbon and hydrogen is decomposed, and yields a mixture of carbon and hydrogen. The hydrogen burns with a colorless flame, but the solid particles of carbon, as they float through the flame, are heated white hot, and furnish the light of the flame. Finally, as these glowing particles reach the outside of the flame, where there is more air, they also burn, evolving still more light.
We are now prepared to study the ways in which such a flame may be converted into what is frequently called a "non-luminous" flame, by which is meant, not a flame that gives out no light, for no such flame exists, but one which emits only the faint light of incandescent gases. For brevity I shall use the term "non-luminous" in this article.
A luminous flame may become "non-luminous" from three causes:
1. Cooling.
2. Dilution of the gas or the air.
3. Too rapid oxidation of the separated carbon.
1. Cooling.—The gas in a flame must be heated to a certain temperature in order that carbon may be separated from it in the manner described.
If the heat of the flame is reduced below this point, no separation of carbon takes place, the flame contains no solid matter, and becomes "non-luminous." For example, if a small gas-flame be caused to play against a cold platinum dish, the flame is spread out over the surface and becomes "non-luminous." The dish, being cold, abstracts so much heat from the flame as to render the separation of carbon impossible. If the dish be heated by a gas-lamp held on the other side, this loss of heat by the flame is checked, and it at once becomes luminous again. A familiar example of the effect of cooling a flame is seen when an ordinary gas-flame is turned very low. Heat is abstracted from the flame by the cooler burner and conducted away, and the result is a small blue flame, emitting scarcely any light. The blue space in the lower part of an ordinary flat gas-flame is likewise due in part to the cooling effect of the burner and in part also to the rush of cold gas into the flame. That this is so is shown by heating the burner, by which means the area of the blue space is notably diminished.
Exact experiments with the photometer have shown that considerably more light is emitted when the gas burns from a red-hot burner, while the consumption of gas is rather diminished than increased. It might seem from these facts that burners of lava, or some similar substance, would have a decided advantage over burners of metal, in regard to the amount of light obtained, since lava is a poor conductor
