Diffraction by a Semi-infinite Screen. 55 



thus a maximum at the surface or the mirror. The last- 

 mentioned result is a direct consequence of: the boundary 



condition |~— =0 I at the surface of the minor assumed by 



Sominerfeld as the basis of his work, but is contrary to the 



observed optical behaviour of any actual screen at very 



oblique incidences. In practice any polished surface is, at 



such incidences, nearly a perfect reflector; but, as shown by 



Lloyd's experiment, both the electric and magnetic vectors 



in tiie incident and reflected waves are of opposite signs at 



the surface irrespective of the plane of polarization of the 



incident light, and their resultant is zero. 



The preceding discussion makes it clear that the solution 



~ds 

 obtained with the boundary condition — - =0 for the li^ht 



vector at the surface of the mirror is entirely inapplicable 

 under experimental conditions for oblique incidences. The 

 solution obtained with the boundary condition ^ = 0) may, 

 however, for our present purpose be regarded as practically 

 valid at all incidences for any screen which is a sufficiently 

 near approach to a perfect reflector, 'provided tJie light be 

 polarized in the plane of incidence. The experimental inves- 

 tigation described in the present paper shows that the same 

 solution may also be regarded as applicable for light polarized 

 at rigid angles to the plane of incidence provided the incidence 

 be very oblique, and attention is confined to the phenomena 

 observed at small angles of diffraction. 



Experimental Methods and Results. 



The diffraction fringes of the Fresnel type bordering the 

 shadow of an obliquely-held screen may be observed with 

 the arrangement shown dia grammatically in fig. la. A 

 front-silvered giass plate bordered by parallel straight edges 

 may be used as the screen. The first edge diffracts the 

 incident light, and by observing the fringes near the surface 

 of the mirror at the second edge, the necessity for using an 

 infinitely extended screen is avoided. The contrast between 

 the maxima and minima of illumination is found to be not 

 entirely independent of the inclination of the surface of the 

 screen to the direction of the incident rays. When the 

 inclination is considerable, the fringes at the edge of the 

 shadow are of course of the usual Fresnel type, few in number 

 and very diffuse. But as the surface is gradually brought up 

 to the position in which it just begins to graze the incident 

 light, the contrast between the maxima and minima of 



