376 



BELL SYSTEM TECHNICAL JOURNAL 



ponent Fnn is again negligible. This behavior indicates that the first 

 term of equation (6) is controlling as the length is increased. 



It appears from all these tests that the magnitude of the far-end 

 crosstalk in this cable with tertiaries terminated does not vary ma- 

 terially from 1500-foot to 12,000-foot cable lengths, except for random 

 effects. In other words, for this range of lengths the second term of 

 (6) is controlling. For very short lengths the crosstalk varies directly 

 with length due to the absence of interaction crosstalk of sufficient mag- 

 nitude to exert any influence. Also, in going from 12,000 to 24,000 

 feet, there is a definite indication that the crosstalk is increasing with 



40 50 60 



80 100 150 200 300 400 500 600 



FREQUENCY IN KILOCYCLES PER SECOND 



600 1000 



Fig. 8 — Crosstalk components in 12,000-foot and 24,000-foot lengths. 



length, so that for lengths over 24,000 feet the crosstalk would again 

 tend to be proportional to length. We have shown in Part I that on 

 the basis of theoretical considerations this law of crosstalk summation 

 with length might be expected. 



To illustrate this measured behavior the far-end crosstalk versus 

 length for frequencies of 50, 100 and 200 kilocycles has been plotted on 

 Fig. 9. For comparison are also plotted dashed curves based on the 

 73-foot tests and computed on the assumption that the crosstalk is 

 directly proportional to length. The difference between correspond- 

 ing curves shows the influence of the tertiary circuits. For a 24,000- 

 foot length this difference amounts to 23, 26 and 27 db at 50, 100 and 

 200 kilocycles, respectively. 



