280 BRUSH— SOME DIFFRACTION PHENOMENA. [April 19. 



hence are seen spread to different extents. This discrepancy is par- 

 tially oft'set by the lateral displacements of the origins due to the 

 curvature of the dift'racting surface, and the net result is that the 

 composite pattern seen is brightest and sharpest in a few fringes 

 only, the position of which may be shifted to some extent by shifting 

 the focal plane. 



Diagram H shows the end of a glass plate with optically plane 

 polished upper surface 12 mm. wide, bounded by straight edges. It 

 may be regarded as a portion of a cylinder of infinite radius, con- 

 stituting one end of a series of curved diffracting surfaces of which 

 the razor edge E is the other limit. The plate is adapted to be 

 slightly rocked by tangent-screw mechanism so that its face may be 

 adjusted very nearly parallel with the incident light. 



When thus adjusted Lloyd's so-called "single-mirror interfer- 

 ence fringes " are brilliantly shown, and the focal plane of the micro- 

 scope may be moved through a wide range over the face of the 

 mirror without disturbing the fringes in any way, proving that they 

 have their origin on the surface of the mirror or plate, and not at 

 its edges. The first one or two dark bands are very black and sharp, 

 and the others show more and more color, until the fifth and beyond 

 are all color. Only seven or eight fringes can be seen, and their 

 spacing is sensibly uniform, as with ordinary interference fringes. 



I shall now endeavor to show that these so-called " single-mirror 

 interference fringes " are not due to interference of light reflected 

 at grazing incidence with contiguous rays not reflected, as commonly 

 supposed, but are superposed diffraction fringes like those already 

 described. 



Considered from this point of view, the origins of the many 

 superposed fringe patterns all lie in the same plane and very nearly 

 in the line of sight, and hence, owing to unequal spreading of the 

 several patterns as already explained, some maxima begin to overlap 

 some minima not far from the major edge of the composite pattern. 

 Therefore few fringes are seen, and most of them are colored. 



The extreme blackness of the dark bands forcibly suggests super- 

 position of many minima. If the very small angle between the face 

 of the mirror and the incident beam of liffht is ^raduallv increased 



