56 



BELL SYSTEM TECJIMCAL JOURNAL 



Hence, if T is the total time of the signal, T is proportional to x^CR 

 (or to KR). That is to say, if tiie duration of the component dots, 

 dashes, spaces of the signal are proportional to the "KR" of the 

 cable, the wave form of the received signal, referred to the r time 

 scale, is invariable, and the total time required to transmit the signal 

 is proportional to the "KR" of the cable. Now the maximum theo- 

 retical speed of transmission on the cable is limited by the require- 

 ment that the received signal shall bear a recognizable likeness to 

 the original system of dots and dashes: in other words there is a 



fit 23456789 10 



Fig. 6 — Elementary telegraph signals in non-inductive cable 



maximum allowable departure in wave form between received and 

 transmitted signals. If, therefore, the actual speeds of two cables 

 are inversely proportional to their "KRs," the wave form will be the 

 same. This establishes Kelvin's "KR" law. As a corollary, if the 

 length of the cable is doubled the speed of signaling is reduced to one- 

 quarter, assuming the same definition of signals. 



The foregoing will be somewhat clearer, perhaps, if we refer to 

 curves, 4, 5, 6, 7 of Fig. 6 which illustrate the distortion sufifered by 

 elementary dot signals in cable transmission. Curve 4 shows the 

 dot signal produced by a unit battery applied to the cable terminals 



for a time inter\al t- 



x-RC 

 2 ' — ; , while curves 5, 6 and 7 are the cor- 



responding dot signals when the battery is applied for the time in- 



x^RC 1 x^RC 1 x^RC 



ter\als ^^ — - — > - " — - — and - - — , — . Any further decrease in the dura- 

 4 2 4 4 4 



tion of the impressed dot, beyond that shown in curve 7, does not 



