ULTRA-SHORT WAVE PROPAGATION 151 



image), but also by negative reflection all along the path. (The term 

 "negative reflection" is used here even in the non-optical case, since 

 when we visualize the process in terms of Huyghen's principle, it is 

 apparent that this case is merely a succession of optical paths.) At 

 higher frequencies this characteristic will cease to rise steadily and at 

 least in the case of simple optical paths will oscillate up and down 

 instead. The rising trend at lower frequencies, however, is found so 

 often that it deserves special mention. It is illustrated by the rising 

 curve and the experimental points shown in Fig. 16. 



A fourth trend is due to diffraction and it is in the same direction as 

 the conductivity trend. Long waves bend more easily about obstacles 

 than do the short; the obstacle may be a mountain or it may be the 

 ever-present bulge of the earth. This type of characteristic is indi- 

 cated in Fig. 16 in the high frequency part of the calculated curve, 

 but in our experiments we have so far not had conditions in which its 

 efi'ect could with certainty be separated from the opposite "negative 

 reflection" trend. The reason for this is that the diffraction trend 

 does not predominate except with frequencies which are sufficiently 

 high. In tests from Deal to Lebanon (Fig. 16) it appears that fre- 

 quencies greater than 1200 megacycles might have to be used in order 

 to separate these effects clearly. This is a point of great importance 

 in view of the wide-spread belief that ultra-short waves suffer most in 

 transmission because of the failure of the waves to bend around 

 obstacles. Except when high mountains or very short waves are 

 involved, the loss in transmission is more likely to be due to re- 

 flection. 



When reflection of vertically polarized waves takes place from a 

 very good conductor, there is no change of phase at reflection, the 

 "negative reflection" mechanism is therefore absent, and the tendency 

 is toward reinforcement rather than cancellation. Physically these 

 conditions can be found in the case of transmission over sea water for 

 frequencies less than 5 mc. In this case, as shown in an as yet un- 

 published study, the diffraction trend has definitely been found experi- 

 mentally and checked quantitatively with theory. 



Optimum Frequencies 



In the preceding pages calculations have been made for various 

 types of path. Both for optical paths and for non-optical paths these 

 have pointed to certain frequencies which, from the transmission 

 standpoint, give most efficient results. The value of this optimum 

 frequency depends almost entirely upon the topography of the path 



