Dust Figures produced by Electric Sparks. 733 



first for interference, it is by no means the only one. For, 

 just as in optics and acoustics, interference may be obtained 

 from a single source by reflexion, that should be possible 

 here also. Indeed, cases of it have already occurred. Thus, 

 the diagonal bars near the corners of figs. 1, 2, and 3, and 

 the central bar in fig. 5 are evidently cases of interference 

 or superposition. 



But this effect is more strikingly shown in fig. 10 by 

 the central bar in the acute angle between two boundaries, 

 which consisted of glass strips on edge. This case may be 

 contrasted with that of fig. 8, in which leaks occurred at 

 the boundaries and no central bar was formed. The diagonal 

 is well shown also at the acute angle in fig. 11 (PI. XXII.). 



Diffraction. — The effects of spreading out beyond openings 

 or behind obstacles was also sought and obtained. Thus in 

 fig. 11 a very small opening in the boundaries was left at 

 the upper left-hand corner, beyond which the spreading of 

 the striae is clearly shown. Also beyond the wide opening 

 at the right the spreading is visible, being specially marked 

 near the two corners. It is also noteworthy that in this 

 right-hand portion the striae are all approximately radial to 

 the spark instead of tangential. This appears to be due to 

 reflexions from the upper and lower boundaries. Thus, the 

 inclined radiation to' the upper boundary gives by reflexion 

 an equally inclined radiation from it; and the resultant effect 

 of these two is equivalent to motions of the air normally to 

 and from the boundary. Hence the formation of striae 

 parallel to that boundary. And this is the case near the 

 boundary in question. Near the opposite boundary the 

 parallelism of the striae to it are also noticeable. 



The series of effects shown in figs. 12-15 were obtained 

 by placing an obstacle of plate-glass 1 inch square in 

 the inverted crystallizing dish previously used for fig. 5. 

 These were planned with a view to securing a spreading 

 behind the obstacle; and they all give evidence of a running 

 of the disturbance along the edge of the square, thus pro- 

 ducing the normal striae with a radial appearance at the 

 corners, 



In each of these four figures (as also in fig. 5) we have 

 portions where the lycopodium is arranged in what may be 

 termed a cellular fashion, i. e., it is specialized though not 

 striated. This distinction may be seen at once by con- 

 trasting this specialized portion with the random distribution 

 of the scattered and unaffected dust outside the crystallizing 

 dish. 



Wldspering Gallery. — Scott Russell has pointed out that, 



