1910] on Electrical and other Pro^ierties of Sand. 751 



resembling that found in the Charlton pits. The length of the one 

 now ready is 3 feet. When the flow begins, a curious rattling sound 

 is heard which finally changes to a distinct musical note. It may be 

 varied slightly, say to the extent of a whole tone or so, by gripping 

 a part of the tube while the sand pours out. The two upper dark 

 bands (Fig. 11) have not become deformed, except slightly at their 

 ends owing to friction between the sand and tube. It is essential 

 for the production of musical sounds, that the ratio of the length of 

 a column to its diameter be such that the upper portion moves down- 

 ward without central deformation. In order to explain the cause of 

 the sound we must therefore consider the motion of this more or less 

 compact body of particles. 



Now, if the lower half of the tube be filled with mercury, and the 

 rest with well-packed sand, the regular lowering of the liquid causes 

 the granular piston to apparently stretch until its extension is about 

 2 per cent, of its original length. It is not until that point is reached 

 that the upper layers begin to move downward. The particles, how- 

 ever, are no longer normally piled. A further slight movement of 

 the lower layers causes the upper ones to follow and to over-run a 

 little (owing to their momentum). Therefore, even if the mercury 

 is adjusted to pour out uniformly from the orifice, the upper part of 

 the sand column moves downward with an intermittent motion, 

 analogous, in fact, to that of a weight drawn over a rough surface by 

 an elastic string. It is also clear that within wide limits the motion 

 of the upper layers may be independent of, or completely out of phase 

 with, that of the lower ones and still produce a musical note. 



The glass wall of the tube is thrown into violent vibration by the 

 intermittent rise and fall of the lateral pressure upon it, so that 

 damping the barrel raises the pitch of the note. The greater part of 

 the sound is due, however, to the direct action of the sand column 

 upon the air above it, for even a tight wrapping of tape but slightly 

 affects its quahty. Where the tube is filled entirely with sand the 

 pitch of the note emitted rises as the column diminishes, owing to a 

 proportional decrease of inertia. 



In order to see in what way varying the friction between the grains 

 would influence the result, we may fill the tube with magnetic sand 

 and magnetise the column longitudinally. This can be conveniently 

 done by winding a current-carrying wire round the tube. With such 

 an arrangement, the sound produced by the descending column, 

 though feeble at first, is strongly increased on magnetising the grains. 

 Each time the circuit is " made," the sound, almost inaudible before, 

 is plainly heard. In all cases the closeness of the grains, i.e. the pro- 

 portion of normally piled particles, largely determines the pitch of the 

 note. Other factors are the state of the glass surface, the size and 

 roughness of the grains, as well as the rate at which they issue from 

 the nozzle. By suitably adjusting all these conditions, a limited 



Vol. XIX. (No. 104) 3 d 



