Experiments icith the Flame Tube. 629 



If one wants to snow the kind of motion during the 

 oscillations of a pendulum, one preferably suspends the 



Fig. 2. 





flame-tube by longer threads, and uses larger amplitudes. 

 Then it is easily seen, that the changes in the sizes of the 

 flames do not take place, when the pendulum is at the point 

 of turnino- back, but when it is near the position of rest, 

 where theacceleration changes sign. 



If another pendulum is hung on to the swinging flame- 

 tube (or vice versa) and the length of both pendulums 

 suitably adjusted, the changes of the flames show how the 

 simple swinging mction is altered by the addition of the 

 octave, twelfth, etc. 



As one can see from the foregoing experiment the flame- 

 tube pendulum might prove of use in detecting recoils. 

 Thus a disk (of cardboard) whose centre of gravity may be 

 displaced from the geometric centre by fastening a small 

 weight to it near the circumference, is fastened to the flame- 

 tube. Rotating the disk (taking care not to swing the 

 flame-tube by the impulse) the flames will remain undisturbed 

 until the weight is brought into play, when the oscillatory 

 reaction becomes noticeable even though the weight may 

 amount to less than one-hundredth of the total weight. 



Similarly the reactionary motion caused by a very small 

 pendulum may be shown. A pendulum consisting of a shot 

 suspended by a thread will be sufficient. 



It can also be shown by means of the flame-tube pendulum, 

 that a system originally at rest cannot alter the position of 

 its centre of gravity, unless acted on by an external force 

 (e. g., case of the recoil of a gun) . For instance, a steel 

 band, the shape of a large tuning-fork, is fastened to the 



