396 



BELL SYSTEM TECHNICAL JOURNAL 



a constant signal meter reading for Test 13. Each horizontal line seg- 

 ment represents one 15-second fluxmeter period (the actual period was 

 14 seconds, one second being required for switching). The arithmetic 

 averages of the segments are shown by dashed lines. It is to be em- 

 phasized that the scattering of improvements taken from adjacent 

 readings is not experimental error but is due to the fact that pairs of 

 adjacent readings were obtained at different stages of the fading cycle. 

 Two separate systems (receivers and antennas and measuring equip- 

 ment) permitting simultaneous measurements would not help this 



TABLE II 



One Branch 



GAU 18,620 kilocycles 



The apparent line loss is 6.7 db (4.3 calculated). 



The equivalent improvement obtained from the data with pads is 7.3 db; without 

 pads, 6.4 db. Average 6.9 or 7 db. 



situation, since fading would not be synchronous on two systems. 

 Evidently a considerable number of readings must be taken to insure 

 that all stages of fading are equally represented in both of the systems 

 being compared. 



Returning to Table II the discrepancy between the apparent line loss 

 and the calculated value (the loss could really be five decibels perhaps) 

 suggests that insufficient data were obtained with the pads. There is 

 a possibility that the measurements without pads involve directional 

 static. However, taking the data as they stand yields an average of 

 seven decibels which is less than one decibel below the value 10 log N 



