As the artificial light of lamps and candles is afforded by the flame they exhibit, it seems a matter of considerable importance to society, to ascertain how the most luminous flame may be produced with the least consumption of combustible matter. There does not appear to be any danger of error in concluding, that the light emitted will be greatest when the matter is completely consumed in the shortest time. It is therefore necessary, that the stream of volatilized combustible gazeous matter should pass into the atmosphere with a certain determinate velocity. If the quantity of this stream should not be duly proportioned; that is to say, if it be too large, its internal parts will not be completely burned for want of contact with the air. If its temperature be below that of ignition, it will not, in many cases, burn when it comes into the open air. And there is a certain velocity at which the quantity of atmospherical air which comes in contact with the vapour will be neither too great nor too small; for too much air will diminish the temperature of the stream of combustible matter so much as very considerably to impede the desired effect, and too little will render the combustion languid. We have an example of a flame too large in the mouths of the chimneys of furnaces, where the luminous part is merely superficial, or of the thickness of about an inch or two, according to circumstances, and the internal part, though hot, will not set fire to paper passed into it through an iron tube; the same defect of air preventing the combustion of the paper, as prevented the interior fluid itself from burning. And in the lamp of Argand we see the advantage of an internal current of air, which renders the combustion perfect by the application of air on both sides of a thin flame. So likewise a small flame is always whiter and more luminous than a larger; and a short snuff of a candle giving out less combustible matter in proportion to the circumambient air; the quantity of light becomes increased to eight or ten times what a long snuff would have afforded. The light of bodies burning with flame, exists previously either combined with the combustible body, or with the substance which supports the combustion. We know that light exists in some bodies as a constituent part, since it is disengaged from them when they enter into new combinations, but we are unable to obtain in a separate state the basis with which it was combined. That in many cases the light evolved by artificial means is derived from the combustible body, is obvious, if we recollect that the colour of the light emitted during the process of combustion varies, and that this variation usually depends not upon the medium which supports the process of combustion, but upon the combustible body itself. Hence the colour of the flame of certain combustibles, even of the purest kind may be tinged by the admixture of various substances.