396 THE FLAME-MANOMETER. 



ing through the india-rubber tube and the tube a into 

 the anterior portion of the capsule, the thin membrane 

 is made to vibrate, these vibrations are communicated 

 to the gas in the posterior part of the capsule, and the 

 gas issues not in a steady flow, but in a series of 

 impulses. Instead of a quietly burning single flame, 

 we obtain a succession of little flames similar to the 

 flame of the chemical harmonicon. Each little flame 

 corresponds to a sonorous vibration, in which the con- 

 densed wave presses the membrane, which thus com- 

 presses the gas behind it and drives a portion of it out 

 of the burner. At the next rarefaction the membrane 

 moves in the opposite direction, the pressure of the gas 

 is diminished, and its efflux either ceases for an instant 

 altogether or becomes at any rate much diminished. 

 If the flame is watched steadily, the eyes being rather 

 close to it, the only effect observable is that it lengthens 

 out in consequence of the action of the sonorous vibra- 

 tions produced by the note which is sounded: the 

 sonorous impulses expel the gas with more force and 

 therefore to a greater height. But the appearance of 

 the flame is very different, if it is observed at a distance 

 of from 1*5 to 3 m , while at the same time the head is 

 moved moderately quickly from side to side, to and fro. 

 In this case the image of the flame impinges upon con- 

 stantly varying portions of the eye, and if a narrow 

 flame is burning steadily, it appears as a broad band, 

 while a flame acted on by sonorous vibrations, by 

 means of our apparatus, appears as a series of little 

 flames placed side by side. 



The motion of the head prevents a close examination 

 of the appearance; besides, it is inconvenient and 



