24 



NATURE 



[July 7, 19 10 



marine cable, where the less the capacity and the less the 

 resistance, the less the time constant, or the quicker the 

 rate of signalling. 



Now the swelling of the pipe or the capacity effect of 

 the cable does not destroy the energy in the water or of 

 the electricity respectively ; this is very different from the 

 waste of energy through resistance, and if by some method 

 we could compensate for the capacity we could signal 



lag) 

 Vo/' 



1714 1714 



Kxcept in Case I. (near its end), the lag in every case is proportional to x 



Kreque- cy, 6"36 per second. 



Submarine telegraph cable -?■= f684 ohms per naut, <i' = o'42 mfd. ye 

 naut. The current received by recorder would be 82 times this if we had n. 

 capacity. 



At -V nauts from sending end the<:e are the voli.s and amperes : — 



I. There is a recorder with 317 ohms resistance at the end of 1825 nau 

 cable 



Infin 



able 



nrys per naut ; no leakance. Not much di^ 

 rjs per naut ; Icakance, 1 '768X10"'' ohms pe 



The above curves are plotted from the resul;>. given in TaMe I. 



through the conductor at any rate we liked, being limited 

 only by the strength of our battery and the sensitiveness 

 of_ our receiver. I may say that the current usually re- 

 ceived would be 1000 times greater if we had no capacity 

 but only the resistance to deal with.' 



As before stated, the cable has resistance ; the current 

 therefore suffers attenuation. It also possesses capacity ; 

 the signalling currents through it therefore suffer distor- 

 tion. Before dealing with this distortion, I must refer 

 you to the diagram of the signals as they are sent into the 

 cable (Fig. 5) and received from it on the siphon recorder. 

 You will notice that the signals, arranged to form thp 

 alphabet in the cable code, are of varying lengths, being 

 I, 2, 3, 4, and 5 times the length of the individual or 

 shortest signal. Sending and receiving on this principle 

 is electrically equivalent to working the cable with varying 

 electrical frequencies of 6, 3, 2, &c., complete periods per 

 second. 



1 I must here refe- to the fact that Mr. Heaviside twenty yeirs ago 

 showed that by giving .series inductance to a cable we could grea 

 our rapidity of sienalling. 



This will be understood from Table I. and Figs. 3 and 4 showi 

 Unfortunately, we see no practical method of carrying out Mr. " 

 suggestion, so that I must go on onsidering the submarine c 



III. Infin 

 tortion. 



IV. Infinite cable. 0-4 h 

 naut to give no distortion. 



(See Figs. 3 .ind 4 ) 



The lower the frequency the less the capacity affects the 

 current, so that the higher frequencies of b and 3 a second 

 are more attenuated than those of 2 and less. The signals 

 that form the letters in the alphabet are differentially 

 attenuated ; the quicker signals, such as those forming a 

 C, are much weaker when they arrive to operate the 

 receiving instrument than the slower signals that form the 

 letters M, O, and so on for the other and longer signals. 



Submarine cable signalling of the present day affords us 

 an electrical illustration of the fable of " the tortoise 

 and the hare " or the principle of " more haste, less 

 speed." 



.As the slower signals get through the cable with more 

 vigour than is necessary, the ingenuity of experimenters is 

 to retard them and to assist as much as possible the 

 quicker ones so that all the signals, whatever their period, 

 shall arrive with exactly the same strength. 



Cromwell Varley in 1862 patented a system for the re- 

 duction of distortion on cables by inserting condensers of 

 suitable capacity in series with the conductor at each end 

 of the cable. 



The reason for the abolition of distortion is obvious ; 

 the condenser absorbs the signals of slow frequency, while 

 the cable transmits them. The condenser allows the 

 signals of high frequency to pass through it, although the 

 cable has attenuated them. It is therefore possible so to 

 arrange the condensers at each end of the line that the 

 condensers and the cable together will more or less correct 

 one another and the distortion be reduced. 



Unfortunately, the absorption of a series condenser is 

 relative, and is inversely proportional to the frequency ; it 

 absorbs more of the slow than the quick signals ; at the 

 same time it does absorb some of the quick, and so far 

 as that is concerned it is harmful ; it diminishes distortion, 

 but at the same time it adds to the attenuation. 



-OATVl/Vn^^. WS^^—^-^ 



allv 



NO. 2123, VOL. 84] 



v^- 



Now " distortion " means something more than the 

 differential transmission of various electrical frequencies ; 

 it also means the " phase relation " of the current to the 

 voltage, and this " phase relation " varies with the various 

 frequencies, so you see that " distortion," looked at from 



