384 Prof. Tyndall on the Action of Sonorous 



hold in my hand a metal box containing compressed air. I 

 turn the cock for a moment, so as to allow a puff to escape : 

 the flame instantly ducks down, not by any transfer of air from 

 the box to the flame ; for I stand at a distance which utterly 

 excludes this idea ; it is the sound that affects the flame. I 

 send a man to the most distant part of the gallery, where he 

 permits the compressed air to issue in puffs from the box ; at 

 every puff the flame suddenly falls. Thus the hiss of the 

 issuing air at the one orifice precipitates the tumult of the 

 flame at the other. 



Finally, I place this musical box upon the table, and permit 

 it to play. The flame behaves like a sentient creature ; bowing 

 slightly to some tones, but courtesying deeply to others. 



I at one time intended to approach this subject of sensitive 

 flames through a series of experiments which, had the flames 

 not been seen, would have appeared more striking than I 

 can expect them to be now. It is not to the flame, as such, 

 that we owe the phenomena which have just been described. 

 Effects substantially the same are obtained when a jet of 

 unignited gas, of carbonic acid, hydrogen, or even air itself, 

 issues from an orifice under proper pressure. None of these 

 gases, however, can be seen in its passage through air; and 

 therefore we must associate with them some substance which, 

 while sharing their motions, will reveal those motions to 

 the eye. The method which we have from time to time em- 

 ployed in this place, of rendering aerial vortices visible, is well 

 known to many of you. By tapping a membrane which closes 

 the broad mouth of a large funnel filled wit^_ smoke, we obtain 

 beautiful smoke-rings, which reveal the motion of the air. By 

 associating smoke with our gas-jets in the present instance, we 

 can also trace their course ; and when this is done, the unignited 

 gas proves as sensitive as the flames. The smoke-jets jump, 

 they shorten, they split into forks, or lengthen into columns 

 when the proper notes are sounded. The experiments are 

 made in this way. Underneath this gasometer are placed two 

 small basins, the one containing hydrochloric acid and the other 

 ammonia. Fumes of sal-ammoniac are thus copiously formed, 

 and mingle with the gas contained in the holder. We may, as 

 already stated, operate with coal-gas, carbonic acid, air, or 

 hydrogen : each of them yields good effects. Here also our ex- 

 cellent steatite burner maintains that supremacy which it exhi- 

 bited with the flames. From it I can cause to issue a thin 

 column of smoke. On sounding the whistle, which was so 

 effective with the flames, it is found ineffective. When, more- 

 over, the highest notes of a series of Pandean pipes are sounded, 

 they are also ineffective. Nor will the lowest notes answer. 



