STUDIES IN RADIO BROADCAST TRANSMISSION 173 



together so we conclude that undulations occur along its surface. 

 These undulations in themselves would cause the length of path of 

 the wave reflected toward the receiver to undergo a continual change. 

 They would also introduce minor reflections from surfaces more 

 distant than that responsible for the major effect which may be 

 responsible for the more rapid, low amplitude fading which is usually 

 superimposed upon the slow changes. Obviously, the character of the 

 fading would in the event that it is caused by undulations along the 

 reflecting layer, be determined by the amplitude and direction of 

 movement over the surface. 



If, on the other hand, we examine the possibilities of theories such 

 as those proposed by Nichols and Schelleng, Larmor and others in 

 which the action of free electrons in the atmosphere is invoked we 

 might visualize the waves on the second path as following a curved 

 trajectory. Or we might have the two sets of waves start off to- 

 gether, become split by double refraction and eventually come to- 

 gether again. Perhaps their planes of polarization will have been 

 rotated. In fact it is possible to build up what appears, we must 

 confess, to be a highly imaginary explanation in which the wave 

 interference is accounted for not on the basis of any great difference 

 in path length but by the assumption that the amount of rotation is 

 such a function of frequency that a change of about 2,000 cycles adds 

 or subtracts a complete rotation, and the further assumption that one 

 set of waves has had its plane of polarization rotated through several 

 more complete rotations than has the other. The synthetic possi- 

 bilities are almost endless and we must wait upon further data more 

 varied in character before the facts can be established. In the present 

 investigation we have not attempted to determine the mechanism of 

 the transmission medium except insofar as it could be inferred from 

 the results of our tests which were aimed at finding out just how radio 

 signals look after they have been subjected to a trip through this 

 mechanism. 



Returning to the solid band fading record illustrated in Fig. 18, 

 let us form some conception of the appearance of this figure were it 

 extended toward the much higher and lower frequencies using as a 

 basis of this conception the supposition that the existing record is 

 systematically distorted by wave interference. For a given rate of 

 change in the physical difference in length of path, such as would be 

 encountered in the simple reflection case, the rate of movement of 

 the minima across the band fading pictures would vary directly with 

 the frequency. Therefore, we can extend the narrow section shown 

 in Fig. 18 to form a wide band fading record such as is shown in 



