734 Prof. E. H. Barton and Mr. W. B. Kilby on 



in the case o£ a water-wave proceeding almost parallel to a 

 wall, reflexion fails to occur ; the line of the wave-crest near 

 the wall setting itself normal to the wall, so that the propa- 

 gation there is along the wall itself. By a similar pheno- 

 menon Lord Rayleigh explains the whispering-gallery effect 

 in sound. Again, in the photography of flying bullets, 

 C. V. Boys, in one case, placed an obstacle to find if this 

 sliding instead of reflexion would occur in the breast and 

 tail waves from a bullet. The photograph showed that 

 it did. 



It accordingly seemed desirable to ascertain if this 

 whispering-gallery effect could be shown by striations of 

 these dust figures. Fig. 16 shows clearly that it can. 

 The arrangements were of the simplest. A curved card was 

 placed on edge as the boundary, the spark occurred on the 

 concave side near one end of it, and a second card, bent 

 sharply at an angle, was placed to protect the curved 

 boundary from any direct effect of the spark. It is seen 

 that normal stria? about an inch long occur all along the 

 interior of this curved boundary, showing that the air 

 motions are along the boundary itself. At the far end of the 

 figure it is noticeable that the effect partly branches off to 

 the other obstacle. 



It is now well to revert to figs. 5, 7, 8, 10, and 12-15. 

 These also show short normal stria? along boundaries curved 

 or straight. The^e striae may possibly be produced by a 

 phenomenon analogous to the whispering-gallery effect. Or, 

 perhaps they may be due to the particular phase-difference 

 between the motions of the air at two given points near the 

 boundary. For, on the theory of Konig *, where the velo- 

 cities differ in phase by more than a quarter of a period, 

 there is attraction between the spheres at the points in 

 question in cases where, with a phase-difference of less than 

 a quarter of a period, there would be repulsion between 

 them. The mechanism of production of some of these 

 normal stria? is accordingly left an open question for the 

 present, but may perhaps repay further investigation. 



Fig. 17 may now be noticed. This shows an effect 

 analogous to that of sound passing through a speaking-tube. 

 Two curved cards were used as boundaries in this case. It 

 may be seen that the disturbance passes freely along the 

 curved space between the cards, also that it spreads a little 

 at the far end. 



Variation of the Pitch of the Strice. — To obtain a striking 

 illustration of the variation of the pitch of the stria? under 

 disturbances of different intensities, sinuous boundaries of 

 * Wied. Ann. xlii. pp. 353, 549 (1891). 



