278 BRUSH— SOME DIFFRACTION PHENOMENA. [April 19, 



der F, of 22 mm. radius. They are very bright and sharp, and nearly 

 free from color. From 12 to 15 may be seen. The, curved plate G, 

 of many cm. radius, gives fringes perceptibly brighter than F. 



When the radius of the cylindrical edge is rather less than one 

 millimeter, all fringes disappear if the focal plane is advanced suffi- 

 ciently to coincide with the median plane of the edge, as would be 

 expected. But when the radius is a millimeter or more, sharp, nar- 

 row fringes may be seen with the focal plane in this position, and 

 these fringes grow broader and more numerous as the radius of the 

 diffracting edge is increased. Evidently they are formed by ele- 

 ments of the cylindrical edge lying beyond (toward the light) the 

 element in the median plane. If, now, the focal plane of the micro- 

 scope is slowly advanced toward the light, these fringes slowly re- 

 treat behind the edge without greatly changing their spacing. They 

 remain visible for some distance behind the edge because the angular 

 aperture of the microscope objective enables the observer to see 

 around and beyond the edge to some extent. Upon reversing the 

 movement of the focal plane the fringes move laterally from behind 

 the edge until the median plane is reached, when the lateral move- 

 ment stops abruptly and the fringe pattern simply broadens out as 

 the retreat of the focal plane continues. 



I am led to the belief that the very greatly enhanced brightness 

 of the fringes produced by the diffracting edge of large radius as 

 compared with the razor edge, is due to the superposition of a 

 number of diffraction fringe patterns which are almost, but not 

 quite, in register. This view is supported by experiments illustrated 

 in diagrams N and O. 



N shows a razor blade greatly enlarged. It makes not the slight- 

 est difference in the fringes whether the blade is in the full line 

 position shown, or in either of the dotted line positions, the essential 

 condition being that the light undergoing diffraction shall not strike 

 the beveled side of the blade. 



At two razor blades are shown clamped together w^ith their 

 edges as close as possible (about 0.2 mm.), and as nearly as possible 

 in the same plane. The combination is adapted to be rotated slightly 

 about the line of one of the edges as an axis by means of a tangent 



