If the flame be extinguished by the breath, and while the wick is smoking the gauze be quickly placed a short distance above it, and a lighted taper applied to the upper side, the ascending combustible gases which still issue from the wick and pass through the gauze will take fire, producing a flame above it, which will not extend beneath because the gauze conducts away the heat sufficiently to prevent ignition of the column of gas below. This phenomenon, however, will only last a moment, as the wick soon ceases, in the absence of heat, to furnish combustible gases. The experiment can be better made over a jet of common illuminating gas. (See fig. 4.) The flame above the gauze will not be so distinctly divided nor so luminous as in an entire flame, because of the partial mixture of oxygen with the combustible gases before passing through the gauze.-If one end of a small glass or metal tube, open at both ends, be introduced into the inner cone of a candle flame, and the other end elevated and a lighted taper applied to it, a second flame will be produced from the combustible gases which have been conveyed off by the tube. (See fig. 5.) It is by the use of such a tube, only longer, and bent so as to pass under water and into collecting vessels, that the gases are collected for analysis.

Bunsen's burner, fig. 6, furnishes an example of the effect of a free and full supply of oxygen to a burning gas. The carbon being consumed almost simultaneously with its hydrogen constituent, scarcely any separation of solid particles occurs, and therefore there is but little light other than that produced by the incandescent gases and vapors. Conversely, the luminosity of a flame may be increased by the addition of substances rich in carbon. If hydrogen gas or light car-buretted hydrogen be passed through naphtha or benzole, its flame may be rendered highly luminous. So also the addition of a substance, as chlorine gas, which has the power of abstracting the constituent hydrogen from a carbo-hydrogen gas and setting free the carbon, will increase the luminosity of a flame.-Increase and diminution of pressure have been found by Frankland to have a remarkable influence upon the luminosity of flames. On the summit of Mont Blanc candles burn with a feeble light, and in artificially rarefied air it has been found that the brightness of ordinary flames increases or diminishes in proportion to the increase or diminution of pressure, down to that which supports a column of mercury of 14 inches. Below this pressure the luminosity diminishes at a less rate than the pressure.

Under increased pressure a flame fed with amylic alcohol was found to increase in direct proportion to the pressure till it was equal to two atmospheres, and beyond this the light increased more rapidly than the pressure. The increase of light is caused by the greater separation of carbon particles under increased pressure, the incandescence of which is the cause of the light. Under a pressure of two atmospheres candle flames evolve much smoke; and the flame of alcohol, which is ordinarily very pale, becomes highly luminous under a pressure of four atmospheres. Conversely, flames which smoke in an ordinary atmosphere cease to do so in a rarefied one, the combustion being more complete in consequence of the greater mobility of the gaseous particles. The reason why the luminosity of flames in very rare atmospheres does not decrease in exact proportion to diminution of pressure is that the incandescent carbon does not furnish all the light; the remainder, which amounts to about 1 per cent. under ordinary circumstances, being produced by incandescent gas, and not being affected by pressure, adding a greater proportional fraction to the amount.-Singing flames were partially investigated by De la Rive in 1802. A small quantity of water heated in the bulb of a thermometer produced musical sounds by the periodic expansion and condensation of vapor in the tube; and he referred the singing of ordinary gas flames in tubes to a similar expansion and condensation of the aqueous vapor formed by the combustion.

Faraday, however, in 1818 showed that flames which did not produce water in burning, such as that of carbonic oxide gas, would produce musical sounds; and that they would also occur in ordinary flames when the surrounding air was raised above 212° F., so that no condensation of vapor could take Place. Experiments in which flames are subjected to the influence of acoustic vibrations producing musical tones show conclusively that the notes produced by them are not of that independent character which would result from expansion and condensation of vapor, but that they have an intimate relation with the principles of harmony. The influence which the length and calibre of the tube in which the combustion takes place, being precisely of the same kind as that exerted on a jet of air blown into an organ pipe, and the sensitive manner in which flames respond to certain musical tones (as has been beautifully illustrated in experiments by Tyndall), indicate their relation to and dependence upon the acoustic vibrations which produce these tones.

This subject, and also that of Konig's sensitive manometric flames, which pulsate on receiving musical vibrations under circumstances in which they indicate by their forms the nature of the sounds, will be treated of in the article Sound.

Flame 700133

Fig. 1.

Flame 700134

Fig. 2. Fig. 3.

Flame 700135

Fig. 4.

Flame 700136

Fig. 5.

Flame 700137

Fig. 6.