136 BELL SYSTEM TECHNICAL JOURNAL 



In most of the cases of radio propagation now being considered, we 

 are concerned with near-grazing incidence since both transmitter and 

 receiver are located near the ground and are separated horizontally by 

 a comparatively large distance. That regular reflection may occur 

 under such circumstances, even over irregular ground, can be shown 

 by a simple optical experiment. A moderately rough piece of paper, 

 such as a sheet of bond or any other paper without gloss is employed. 

 The paper on which this is printed is rather too smooth to give a 

 striking result, but it may be used. If the reader will focus his eye 

 on some distant object which shows up with contrast against the sky, 

 and if he will then hold the paper about a foot from the eye so that the 

 line of sight is parallel and very close to the plane of the paper, it will 

 be seen that the rough sheet has become a surface with a high gloss. 

 It is helpful to bend the paper slightly so as to produce a cylindrical 

 surface having elements parallel to the line of sight. Images of 



77777777777777777777^ 



Fig. 9 



distant objects can be seen clearly in such a paper mirror and con- 

 siderable detail can be obtained provided that the angle of incidence 

 differs from 90° by something less than one degree. It is to be re- 

 membered that in most of the optical paths encountered in ultra- 

 short wave propagation, we are concerned with angles which are as 

 near to grazing as this is. 



The reason for this reflection from a rough surface is readily explained 

 on the basis of Huyghens' principle. The situation is represented in 

 Fig. 9. Let us suppose that the general level of the rough surface is 

 below the line of sight TOR by a distance H. H is assumed small 

 compared with D, the length of the path. As a result of variations in 

 H due to the ruggedness of the terrain there will be corresponding 

 variations in the total length of the optical path TSR. Reflections will 

 be approximately regular, however, if these variations in TSR are 

 small enough in comparison with half a wave-length. In Fig. 9, a 

 change in level,. /?, is represented at S, the dotted line representing an 

 irregularity which has been added. These assumptions lead readily 



