1192 THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER, 1953 



curves impractical without a computer capable of handling a relatively- 

 large number of components several of which may be relatively small. 



However, experimentation with a trial and error method of solution 

 on a few selected curves has shed considerable light on the nature of 

 received signals which could produce the types of curves included in 

 the data under consideration. 



Curve (a) on Fig. 4 is one of those selected for analysis and was syn- 

 thesized by using a direct signal component and six interfering com- 

 ponents as shown in Table I. If components 5 and 6 had been omitted 

 from the synthesis, curve (b) would have resulted; and, if components, 

 3, 4, 5 and 6 had been omitted curve (c) would have been obtained. 



The difference between curves (a) and (b) illustrate the radical effect 

 which can be produced on the shape of a curve near maxima of path loss 



10 



20 



30 



-J 

 < 



i 40 

 O 



9 



g 10 



o 



$30 



z 



^40 

 8 



5 50 

 a. 

 



30 



3800 3900 4000 4100 



FREQUENCY IN MEGACYCLES PER SECOND 



Fig. 2 — Typical path-loss versus frequency curves observed on the Lowden- 

 Princeton, Iowa, path during July and August, 1960. 



