H. V. Gill — Theory of Singing Flames. 185 



be sounded by a siren or other means. He says that this is an 

 example of the propagation of sound, by vibrations through 

 the air, and its reception by a body extremely sensitive to such 

 influences. If the flame is smaller, it continues singing once it 

 has begun to respond. That is to say, the flame takes up the 

 vibrations of the air inside the tube, reacts on them in the way 

 we have described, and thus the sound is strengthened into the 

 continuous note with which we are familiar. If we examine 

 the flame in a rotating mirror before it has begun to sing, we 

 see of course a continuous band of light. Now, if we sound 

 the note of the tube, regarding the flame in the rotating mir- 

 ror while doing so, we see the image as represented in tig. 6. 



/// 



From this we see clearly the mutual reaction between the 

 flame and the column of air. If the flame be so small that it 

 is just on the point of beginning to sing of itself, a sudden 

 noise, such as clapping the hands, will cause it to begin. 

 Always, however, the gradual development as shown in tig. 6 

 tajses place. Again, a small flame may be made to sing by 

 blowing gently across the top of the tube, and as before shows 

 the gradual development. These facts are sufficient to show 

 how the pressure due to the vibration of the column of air and 

 the pressure of the gas react on each other. It is scarcely neces- 

 sary to add that in reality this " gradual development " takes 

 place in a very short interval of time, especially in the case of 

 a small plane inside a short tube. 



From what we have seen we are led to the conclusion that, 

 when the flame begins spontaneously, in reality the cause is to 

 be found in some change of pressure taking place in the column 

 of air. If we slowly lessen the flame, before it begins to sing, 

 regarding it as before in the rotating mirror, we notice, at a 

 certain point, that gentle undulations appear on the border of 

 the band of light, which follow the cause of gradual develop- 

 ment above described. Is there any such external cause at 

 work ? We think we shall be able to show that there is. We 

 have seen that the slight changes of pressure due to a note 

 sounded at some distance from the tube produce changes in the 

 pressure sufficient to cause the flame inside the tube to sing 

 (Schaffgotsch put a small flame into action by the noise caused 



