1194 THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER, 1953 



_i 

 < 



a. 

 o 



z 



r 



a 



q: 



-I 



^ 15 



O 



UJ 



o 



z 



{^520 

 O 



25 

 3750 



3800 



3850 3900 3950 4000 4050 



FREQUENCY IN MEGACYCLES PER SECOND 



4100 



4150 



Fig. 4. — Path-loss versus frequency curves synthesized from components 

 shown in Table I. 



selectivity. (Statistical data are given later which bear out this 

 conclusion). 



2. No deep fades were found that did not include an appreciable 

 component of loss with little appearance of frequency selectivity; that 

 is, no deep fading occurred unless the signal was already depressed 10 

 or so db across the entire observed band. This could be caused either 

 by (a) the presence of appreciable interfering signals of slight delay from 

 the direct signal component, see curve (c) on Fig. 4, or (b) by non-fre- 

 quency selective attenuation of the direct signal component. The former 

 seems more likely; but the evidence is not conclusive in this regard. 



3. No curves for deep fades were found which appeared as if they could 

 be synthesized satisfactorily with fewer than four to six components. 



Table I 



