■9-3.1 BRUSH— SOME DIFFRACTION PHENOMENA. 279 



screw, so that tlie edge nearer the light may be withdrawn very 

 slightly below the plane of the incident beam which strikes the other 

 edge. When this adjustment is just right the brightness of the one- 

 blade fringes is approximately doubled, clearly indicating that two 

 superposed fringe-patterns are formed. It appears that twice as 

 many elements of each wave front are affected. 



We may regard the cylindrical dift'racting surfaces as consisting 

 of a great many parallel elements, each acting as a diffracting edge 

 and producing its own fringe pattern which is superposed on those 

 of the other elements. This superposition of fringes is not apparent 

 when they are viewed in the usual way, /. c, in a plane far removed 

 from the diffracting edge, because nearly all of the patterns have 

 their origins so far behind (toward the light) the tangent element of 

 the edge that they are hidden by it. The method of viewing the 

 fringes herein described, however, enables the observer to see these 

 hidden fringe patterns, as already pointed out. 



Measurements, the details of which need not be gone into, show 

 that in the case of the cylinder F, of 22 mm. radius, the width of 

 the strip of surface involved in producing the best and brightest 

 fringe pattern is about 1.5 mm., though 0.9 mm. gives all but the 

 extreme lines. Smoking the surface of the cylinder makes very little 

 difference in the brightness of the fringes, and the slight loss ob- 

 served is accounted for by the roughening of the surface. 



Careful eyepiece micrometer measurements of the spacing of the 

 fringes formed by the razor edge E, and a cylinder of small radius 

 agree perfectly with the theoretical spacing of diffraction fringes. 

 But with the large cylinder F (and still more so with the curved 

 surface G) the spaces diminish less rapidly toward the outer margin 

 of the pattern and the outer fringes lose their sharpness, because 

 the many superposed fringe patterns which form the composite pat- 

 tern observed are not quite in register; so that beyond 12 or 15 

 fringes many maxima and minima so far coincide that no more lines 

 are seen. 



The reason why the numerous patterns are not perfectly in reg- 

 ister becomes clear when we consider that they have their origins 

 at different distances from the focal plane of the microscope, and 



