74 BELL SYSTEM TECHNICAL JOURNAL 



the small irregularities which exist in the make-up of any practical 

 loaded line. A further cause is the reflection at the adjacent repeat- 

 ers, due to the difference between the repeater impedance and the 

 line impedance. 



It will be noted that three sets of echoes are shown which affect 

 the " talker". In addition to these which involve one or more repeat- 

 ers, a comparatively small amount of power is reflected back to the 

 "talker" from the various irregularities between the "talker" station 

 and the nearest repeater. These reflections have not been indicated 

 since their effects are of negligible importance. Six sets of echoes 

 affect the "listener". Both for the echoes affecting the "talker" and 

 the "listener", the dotted lines indicate reflections from a number 

 of different points where irregularities exist as explained above. 



In circuits containing a larger number of repeaters the numbers 



of sets of echoes affecting the talker and listener are, of course, greater. 



The number of sets of first echoes affecting the talker is equal to the 



number of repeaters. The number affecting the listener is equal to 



N(N+1) AT . , , 



where A is the number ot repeaters. 



It is, of course, obvious, that, for either four-wire or two-wire 

 circuits, if the circulating energies are large, they will have an 

 adverse effect on the ability of two people to carry on a conversation 

 over a telephone circuit. Xot only will the transmission received by 

 the listener be adversely affected, but the talker will be considerably 

 distracted, particularly when the time of the transmission over the 

 circuit is so long that he hears a distinct echo of his words. 



Experiments have shown that the effects of the echoes both on 

 the listener and talker become more serious as their time lag is in- 

 creased. This means that as telephone circuits are made longer it 

 is necessary either to improve balances or to design the telephone 

 circuits so that the velocity of propagation will be higher. This 

 necessity for making the velocity of propagation high on long cir- 

 cuits was one of the principal reasons which led to the selection of 

 extra light loading for the longer circuits. 



Figure 5 shows very approximately how the effects of the echoes 

 vary with the length of time by which they are delayed. One curve 

 is given for the effect on the "talker", another for the effect on the 

 "listener". Both curves indicate, for various time lags, the com- 

 parative magnitude of echoes which are small enough to be inappre- 

 ciable when ordinary telephone conversations are carried on. The 

 curve applying to the "listener" is referred to the direct power which 



