

Mr. F. Guthrie on Approach caused by Vibration. 349 



thest from the fork, fig. 6 (1) and (2). When both plate and 

 fork were accurately horizontal, and the strongly vibrating fork 

 was very close to the powder, the latter partly separated into 

 irregularly shaped flakes which showed no inclination to adjust 

 themselves into any definite geometric form. 



§ 15. It may be remarked that when powder is strewn on the 

 face b, when the fork is vibrating as in § 14, the particles are 

 jerked in straight lines backwards and forwards to distances de- 

 pending upon their masses. These lines cross the central line 

 of the fork's Z>~face; and although, after a time, it is along this 

 line that powder chiefly remains, this is because, at other parts, 

 the powder is cast off at one or other edge as each particle falls 

 within the amplitude of the fork's vibrations. 



§ 16. I conclude from §§ 14< and 15 that no sensible whirl- 

 winds are established in the air in contact with the Z>-face when 

 the plane of vibration is parallel to that face. 



§ 17. Experiment 6. — The fork was fastened vertically down- 

 wards (Hg); and being set in vibration as in § 11, the glass plate 

 strewn with magnesia was approached vertically. As in Experi- 

 ment 5, the two had to be brought very near before any disturb- 

 ance occurred. When within 1 or 2 millims. the magnesia began 

 to be scattered, the portion nearest to the edges parallel to / 

 (fig. 1) being most disturbed and being driven from under the 

 fork. 



This distribution of the powder is evidently to be traced to the 

 different lengths of path and consequent different rates of motion 

 of the different parts of the face c. The edges parallel to / (fig. 1 ) 

 move further and faster than the central line between them, since 

 (fig. 8) C A > C B. Indeed, assuming that the prong remains 

 unbent and swings as a pendulum about 0, then 



mean velocity of A __ C A 

 mean velocity of B ~" C B ' 

 or in the case of fork A, which has the dimensions given in § 4, 



mean velocity of A 3256 _ ~ nrin 



— J — — = — — - = 1 0003. 



mean velocity ot B 3255 



The result is that the prevailing tendency to disturbance de- 

 pends upon the interference of two currents, p and q (fig. 8), which 

 are set up alternately. As before, the space v between p and q 

 willbe a spot of comparative calm, but not a spot for accumulation. 



§ 18. If the face c be horizontal but turned upwards, and if 

 the powder be sprinkled upon the upturned face, it is set in vio- 

 lent agitation. Each particle moves to and fro with great velocity 

 in the plane of vibration. Those nearest the edge are pretty sure 

 to be presently thrown off. Those which are so situated as to be 



