NAIL'UAL rillLOsOlMIY. 245 



where it will cease altogether. Above this point, for a certain distance, the 

 flame may be caused to burn tranquilly and silently for any length of time, 

 but when excited by the voice it will sing. 



When the flame is too near the point (A), on being excited by the voice or 

 by a tuning-fork, it will respond i'c: a short time, and then cease. A little 

 above the point where this cessation occurs, the flame burns tranquilly, if 

 unexcited, but if once caused to sing it will continue to sing. With such a 

 flame, which is not too sensitive to external impressions, I have been ublo 

 to reverse the effect hitherto described, and to stop the song at pleasure by the 

 sound of my voice, or by a tuning-fork, without quenching the flame itself. 

 Such a flame I iintl may be made to obey the Avord of command, and to 

 sing or cease to sing as the experimenter pleases. 



The mere clapping of the hands, producing an explosion, shouting at an 

 incorrect pitch, shaking of the tube surrounding the flame, are, when the ar- 

 rangements are properly made, ineffectual. Each of these modes of dis- 

 turbance doubtless affects the flame, but the impulses do not accumulate, 

 as in the case where the note of the tube itself is struck. It appears as if 

 the flame were deaf to a single impulse, as the tympanum would probably 

 be, and, like the latter, needs the accumulation of impulses to give it suf- 

 ficient motion. A difference of half a tone between two tuning-forks is suf- 

 ficient to cause one of these to set the flame singing, while the other is power- 

 less to produce this effect. 



I have said that the voice must be pitched to the note of the tube which 

 surrounds the flame ; it would be more correct to say the note produced by 

 the flame when singing. In all cases, this note is sensibly higher than that 

 due to the open tube which surrounds the flame ; this ought to be the case, 

 because of the high temperature of the vibrating column. An open tube, 

 for example, which, when a tuning-fork is held over its end, gives a maxi- 

 mum reinforcement, produces, when surrounding a single flame, a note 

 higher than that of the fork. To obtain the latter note, the tube must be 

 sensibly longer. 



What is the constitution of the flame of gas while it pi-oduces musical 

 sounds ? This is the next question to which I will briefly call attention. 

 Looked at with the naked eye, the sounding flame appears constant ; but is 

 the constancy real 1 Supposing each pulse to be accompanied by a physical 

 change of the flame, such a change would not be perceptible to the naked 

 eye, on account of the velocity with which the pulses succeed -each oilier. 

 The light of the flame would appear continuous on the same principle that 

 the troubled portion of a descending liquid jet appears continuous, although 

 by proper means this portion of a jet can be shown to be composed of iso- 

 lated drops. If we cause the image of the flame to pass speedily over dif- 

 ferent portions of the retina, the changes accompanying the periodic im- 

 pulses will manifest themselves in the character of the image thus traced. 



I took a glass tube, three feet two inches long, and about an inch and a 

 half in internal diameter, and, placing it over a very small flame of defiant 

 gas (common gas will also answer), obtained the fundamental note of the 

 tube. On moving the head to and fro, the image of the sounding flame was 



21* 



