10 W. B. Rogers on Sonorous Flames. 
it turns freely. This, which is represented in the figure by the 
dotted lines, has two little cushions of buckskin glued on the 
top, for the convenience - supporting the larger tubes during 
the experiment. 
making this chpoeaar? it will be remarked that although 
in vit first instance the rotation of the flame often serves to ex- 
cite the vibration, an increase of velocity eukeus the force of 
the sound, and a very high speed actually annuls it, the flame 
at the same time+resuming its normal appearance of continuity. 
This effect is not due, as might be supposed, to the air carried = 
round by the revolving motion obstructing the full supply of 
oxygen to the flame, for I have found that it is not perceptibly =| 
increased by attaching a disc of card to the revolving arm of the 
nen and moreover I have observed that a small singing 
r being silenced by very ae rotation shows through- 
out its circuit the blue column due to a full admixture of air. 
More probably the cause of this ay oe and destruction of 
the sound should be sought for in the mechanical interference of 
the successive pulses due to explosions at widely separated points 
in the orbit of the flame; but further observations are needed 
clearly to explain the effect. 
_ As in the explosive state the flame has extremely little illu- 
minati wer, we must conclude that in the above as well as 
the following experiments, the obscure intervals mark the periods 
of explosion, and the bright ones the periods in which the gas 
burns in the ordinary way with a luminous white flame. 
5. The intermissions of the sin ong. fee are made apparent oy a 
er Te 
flame between the successive explosions makes this part of the 
pipe strongly vale it assumes the aspect of a number of 
: to the subdivisions of the crown of flame; 
and if to vary tbs: effect we blacken the horizontal arm and 
fasten near its outer end, or where it er the vertical direc- 
tion, a brilliant bead of glass or metal, we are presented with a 
circle of starry points eee of which by a proper adjustment of 
motion appears to rest. 
following proof of “2 intermitting nature of the singing 
2, Suggested by the effects just described, is at once so simple 
‘Teadil y peen at a distance, as perhaps to merit a place 
I lecture-room illustrations. In this experiment the 
e flame is held at rest in the tube and the 
| by receiving the light on a circular 
