204 Mr. 0. Heaviside on Electromagnetic Waves, and the 



or the speed of the waves is v, and the attenuating coefficient 

 P is practically independent of the frequency, and is made 

 smaller by reducing the resistance and increasing the induc- 

 tance, of the dielectric. 



The corresponding current is 



C=Y/L v 



very nearly, or V and C are nearly in the same phase, like 

 undissipated plane waves. There is very little distortion in 

 transit. 



How to increase L is to separate the conductors, if twin 

 wires, or raise the wire higher from the ground, if a single 

 wire with earth-return. It is not, however, to be concluded 

 that L could be increased indefinitely with advantage. If I 

 is the length of the circuit, 



m = 2L v 

 shows the value of L which makes the received current 

 greatest. It is then far greater than is practically wanted, so 

 that the difficulty of increasing L sufficiently is counter- 

 balanced by the non-necessity. The best value of L is, in the 

 case of a long line, out of reach ; so that we may say, gene- 

 rally, that increasing the inductance is always of advantage to 

 reduce the attenuation and the distortion. 



Now if we introduce leakage, such that R/L n = K/S, we 

 entirely remove the distortion, not merely when R/L nis small 

 but of any sort of waves. It is, however, at the expense of 

 increased attenuation. The condition of greatest received 

 current, L being variable, is now 



We have thus two ways of securing good transmission of 

 electromagnetic waves : one very perfect, for any kind of 

 signals ; the other less perfect, and limited to the case of 

 R/L n small, but quite practical. The next step is to secure 

 that the receiving-instrument shall not introduce further dis- 

 tortion by the quasi-resonance that occurs. In the truly 

 non-distortional circuit this can be done by making the resist- 

 ance of the receiver to be L Q v (whatever the length of the 

 line) ; this causes complete absorption of the arriving waxes. 

 In the other case, of R/Ln small, with good insulation, we 

 require the resistance of the receiver to be L t> to secure this 

 result approximately. I have also found that this value of the 

 receiver's resistance is exactly the one that (when size of wire 

 in receiver is variable) makes the magnetic force, and therefore 

 the strength of signal, a maximum. Some correction is 

 required on account of the self-induction of the receiver ; 



