184: H. V. Gill — Theory of Singing Flames. 



this is so, for the size of the flame depends on the pressure of 

 the gas which feeds it. We see also why a gas-generating 

 apparatus is disadvantageous, because, as we have before re- 

 marked, it is impossible to regulate the gas pressure with suffi- 

 cient exactness. With the flask we have above described, we 

 were able to see that the notes produced depended in great 

 measure on the pressure, and that also the intensity of the 

 sound depended on the same cause. We also noticed that, 

 immediately on the note beginning to sing, the water rose a 

 few mm. in the pressure tube, thus showing that less gas was 

 used when the gas was singing than when silent, a fact which 

 was also noticed by Count Schaffgotsch, from a different ex- 

 periment. 



Thus far we have considered a flame singing under the most 

 favorable conditions. Often, however, if the size of the flame 

 (i. e., the pressure of the gas) and its position in the tube be 

 carefully regulated, the note produced, though continuous, is 

 faint, and does not develop into a full, loud sound. The image 

 of the flame in this case does not present a series of distinct 

 tongues as in the former case, but resembles that of a mano- 

 metric flame. On considering what we have hitherto said, 

 this case presents no difficulty. It is only necessary to remark 

 that the pressure of the gas and its position being such, the 

 reactions just described are not sufficiently marked to pro- 

 duce the full result. If the flame be distant from the node to 

 which its size is well adapted, it is clear, since the changes of 

 pressure are less intense the farther we go from the node, that 

 the flame cannot sing as it would do if under more favorable 

 conditions. In the same way is to be explained the fact that 

 the note is not always that which would require a node at that 

 part of the tube where the flame is placed, for it may happen 

 that the size of the flame corresponds better to the period of a 

 note which has a node farther from it, which note it will rein- 

 force rather than the one which has a node nearer to it. So 

 also is easily explained why the flame may emit two notes 

 simultaneously — i. e., if it be placed so as to be near both nodes. 

 It may be said that the limit of pressure required for a given 

 note is fairly wide. 



We think we have thus sufficiently shown the importance of 

 the pressure on the gas in maintaining the sound. We have 

 now to show how the flame begins to sing. The flame may 

 begin to sing without any apparent external cause, or may 

 be put into action by an external cause. We hope to 

 show these two cases may be reduced to the latter. Tyndall, 

 in his usual graphic way, describes in his " Heat " how a 

 flame, too large to sing continuously, will respond to the 

 note proper to the tube inside which it is placed, if this note 



