W. B. Rogers on Sonorous Flames. 11 
may be revolved by the impulse of the hand. The face of the 
disc next the tube, colored of a barre nie with paint or a cov- 
ering of cloth, should have a nar strip of white paper 
fastened upon it ina radial dinedtagt: ie small circular bit of 
the same placed near the edge. If bo th faces are used alternately 
we may affix the white bar to one and thi ‘dot to the other. On 
bringing the six-feet tube down over the flame and giving rapid 
motion to the disc, we remark that so long as the flame contin- 
ues silent the bar or dot is omy invisible, but as soon as the 
sound commences the black disk becomes diversified by a series 
of whitish images of one or other of these objects arranged at 
equal ‘intervals around the central point. Itis scarcely neces- 
sary to say that the number of these images as well as their 
apparent motion and rest will depend on the time of rotation as 
compared with the interval of the explosions of the flame. 
Should it happen that the period of one revolution of the wheel 
is precisely that of a certain number of the explosions neither 
more nor less or that of any multiple or sub-multiple of this 
number, the images of the bar or dot will continue in each suc- 
cessive rotation to present themselves at the same points, but 
should such a relation not subsist these images will be seen to 
shift their places on the disc sometimes appearing to advance and 
at others to retreat. 
6. On the primary source of the vibration of singing flames.— 
The ala experiments taken in connection with those of 
co 
ures i ) : 
experiments of Savart* nage ‘inces ps Aaa the pabceate! re- 
searches of Masson+ and Sondbaucel have furnished abundant 
* It pears that these results announced by Savart in his Lectures on Acoustics 
in ar _— e de France were published by Masson i ves the Journal d’Institut. 
fisskes de Chimie et de Physique, 3e Ser., 
% Annalen der Physik und Chemie. Band xci. 
