smaller portion; these increase with the sound, which at last becomes very loud, and if the flame be further introduced into the tube, it is generally blown out. Such are the general appearances with hydrogen. If a jet of olefant or coal-gas, both of which I have ascertained may be used successfully, be substituted, then, in addition to those appearances, it will be perceived, that as the bright flame of the gas enters the tube, its splendour is diminished, and it burns with less light. By substituting other gases and inflammable vapours for hydrogen, and using other vessels than tubes, I was enabled so to magnify the effects, as to perceive more distinctly what took place in the flame at these times, and soon concluded that the sound was nothing more than the report of a continued explosion.
Sir H. Davy has explained the nature of flame perfectly; and has shown that it is always a combination of the elements of explosive atmospheres. In continued flame, as of a jet of ga, the combination takes place successively, and without noise, as the explosive mixture is made. In what is properly called an explosion, the combination takes place at once throughout a considerable quantity of mixture, and sound results from the mechanical forces thus suddenly brought into action; and a roaring flame presents some of the characters of both. If a strong flame be blown on by the mouth, a pair of bellows, the draught of a chimney, or other means, the air and the gaseous inflammable matter are made to mix in explosive proportions in considerable quantities at once, and these being fired by the contiguous flame, combine at once throughout their whole extent, and produce sound: the effect is rapidly repeated in various parts of the flame as long as the air is mixed thus forcibly with it, and a repetition of noise is produced, which constitutes the roar.
Now this I believe to be exactly analogous to that which takes place in what have been called the singing tubes; but in them the explosions are generally more minute and more rapid. By placing the flame in the tube, a strong current of air is determined up it, which envelopes the flame on every side. The current is stronger in the axis of the tube than in any other part, in consequence of the friction at the sides and the position of the flame in the middle; and just at the end
