138 



Profs. W. E. Ayrton and John Perry. [Dec. 12, 



the point S. But if we now turn to fig. 3, where the light is con- 

 verging to a point behind the convex surface, and nearer to the sur- 

 face than half the radius of the mirror, then, after reflection, the light 

 converges to a point in front of the mirror, and, as before, the area 

 GH (which has become almost a point, G) is smaller, and therefore 

 brighter, than the area DE, as long as the screen is nearer to the mirror 

 than the point P, but larger, and therefore darker, than D'E', when 

 the screen is farther from the mirror than P. In other words, if the 

 phenomenon of the Japanese mirror is due to the curvature of different 

 parts of the surface being slightly different, then with the arrange- 



Fig. 3. 



ment of light shown in fig. 3, the whole effect ought to undergo an 

 inversion as the screen passes through P ; that is to say, if the parts 

 corresponding with the pattern at the back are the flatter, then, while 

 these should appear as bright on a dark ground when the screen is at 

 a position DP, they ought to appear as dark on a bright ground when 

 the screen is at a position D'F'. Now this is exactly what is found to 

 be the case when tested experimentally. 



Again, if the phenomenon is, as the previous experiment would lead 

 us to conclude, due not to unequal reflecting power of the different 

 portions of the surface of the mirror, but to minute inequalities on 

 the surface, in consequence of which there is more scattering of the 

 rays of light falling on one portion than on another, then since rays of 

 light making very small angles with one another do not separate per- 

 ceptibly until they have gone some distance, it follows, that if the 



