NATURAL PHILOSOPHY. 187 



flame by simply moving it to and fro without striking the tube, cannot be 

 thus explained, since the gentle impulse given to the air by the vibrating 

 pipe produces no audible effect, and would seem quite inadequate to excite in 

 the column any but the very feeblest vibration. Admitting, however, that 

 these extremely faint vibrations of the air may contribute somewhat to the 

 result, it can hardly be doubted that the main influence by which the move- 

 ment of the jet produces the effect in question is by causing so rapid a mix- 

 ture of the adjoining air with the gas that the latter, before being inflame.!, 

 is brought into a condition to produce those small explosions, which, by 

 their quick succession, give origin to the sound. The effect of motion in 

 bringing about this explosive condition of the flame is well exemplified by 

 the following experiment : 



(2.) Fastening a jet-pipe, some ten inches in length into the end of the 

 flexible tube through which the gas is supplied, and holding it erect by a 

 point a little below the insertion so that we can readily cause it to vibrate, 

 we ignite the gas and adjust it to form a slender flame about an inch long. 

 If now the flame be moved from side to side at a moderately rapid rate, it 

 will assume, according to a well-known visual law, the appearance of a con- 

 tinuous arch of whitish light, retaining at the extremities the whole height of 

 the stationary flame, but growing narrower from either side towards the 

 middle. In these conditions the flame is entirely noiseless. As we gradually 

 increase the speed of the vibrations, the arch, at a certain stage, suddenly 

 breaks in the middle, where a faint bluish flame takes the place of the usual 

 whitish light, and at the same instant a sharp noise is heard, due to the inflam- 

 mation of the explosive mixture at this part of the vibration. It is hardly 

 necessary to say that as the vibration is quickest at the midway point, com- 

 ing to a pause at each end of the arch, the gas becomes more largely mixed 

 with air at the middle than towards the extremities of the motion, and is, 

 therefore, at this point earliest reduced to the state of an explosive mixture. 



As we increase the velocity of the vibrations, the sonorous part of the arch 

 extends towards the ends, until the path of the flame presents the aspect of 

 a narrow, bluish band, irregularly serrated at the top, and flanked at the ex- 

 tremities by a tall flame of the usual whitish color. As might be expected, 

 when the jet is revolved rapidly in a circle, the white light entirely disap- 

 pears, and the ring of bluish flame which results gives forth a continuous 

 but not distinctly musical sound. When made in a dark room these simple 

 experiments were found to be unexpectedly interesting and beautiful. 



(3.) As in the above cases the action is mainly traceable to the more rapid 

 mingling of atmospheric air with the flame, it is natural to conclude that a 

 like effect would be produced bypassing a current of air upwards through 

 the tube, and on trial this anticipation was strikingly verified. In order that 

 the current may be evenly distributed, it is convenient to employ an argand 

 burner, having the supply-pipe at the side, and the central opening entirely 

 free, so that the jet-pipe may rise through the centre, and the burner be 

 adjusted to the proper distance below the flame and the bottom of the glass 

 tube. The air conveyed to the argand burner through a flexible pipe, may 

 be supplied either from the lungs of the operator or from an adjacent gas- 

 ometer. In most cases the action of the current is more easily managed 

 when the apertures from which it flows are some two or three inches below 

 the bottom of the resonant glass tube. 



"With this arrangement, and a proper graduation of the current of air 

 impelled into the tube, we can cause the flame to sing when the other meth- 



