72 THE BELL SYSTEM TECHNICAL JOI'RNAL, JANUARY 1952 



no significant part in the fading picture with the exception of the situa- 

 tion illustrated in Fig. 3(c). Occasionally on the Murray Hill path, 

 conditions of atmospheric refraction are such that a strong signal com- 

 ponent is received by virtue of reflection from the water surface of 

 Raritan Bay. Under normal conditions, the geometry of the path does 

 not permit such a reflection. 



Normally the dielectric constant of the atmosphere decreases with 

 height above ground so that the ray path usually has a curvature in 

 the same direction as the earth cur^^ature. However, it is possible for the 

 dielectric constant of the atmosphere to increase with height above 

 ground (sub-standard conditions) so that the ray path has a curvature 

 opposite that of the earth. This results in the condition illustrated in 

 Fig. 3(d) where the limiting or tangent ray does not reach the receiver 

 and only a weak signal is received by virtue of diffraction. Widely sepa- 

 rated frequencies and vertically spaced antennas are affected alike as 

 regards the average signal level but not the fine structure fading. This 

 effect has been observed only on the Southard Hill-Crawford Hill path 

 which has small clearance to begin with. It has been observed on several 

 nights in late summer or early autumn after a radiation type ground 

 fog has formed in the late evening and usually persists until the fog is 

 dispelled by winds or by the morning sun. 



There are, of course, times when the transmission conditions are con- 

 siderably more complicated than those described above. Some of these 

 apparently are due to a combination of the situations illustrated in Fig. 3 

 while others may be the result of an atmospheric focussing or trapping 

 phenomenon. In addition to the various phenomena just described, which, 

 fortunately, occur rather infrequently, there are numberous occasions 

 Avhen the signal varies plus and minus a few decibels relative to the free 

 space level. It has not been possible actually to demonstrate the mecha- 

 nism responsible but it seems most likely that these smaller variations 

 are due to non-linear dielectric constant gradients which give the atmos- 

 phere the properties of a convergent or divergent lens. 



An important result of the observations made to date is the con- 

 viction that the severe fades, signal excursions to levels 30 decibels or 

 more below the free space field, were all caused by wave interference. It 

 appears that, as the average signal level is depres.sed by any mechanism, 

 it becomes more and more vulnerable to the effects of extra signal 

 components of small amplitude that often may be present but go un- 

 noticed when the signal is near normal levels. Thus, while the average 

 signal level during the conditions illustrated in Figs. 3(b) and 3(d) may 

 be no more than 15 to 20 decibels below the normal daytime level, there 



