1914] on Improvements in Long-Distance Telephony 93 



of the electromotive force at the sending of the cable. The other 

 wire can be employed to pick up and depict the wave form at any 

 required distance, 10 miles, 20 miles, and so on up to the full length 

 of loO miles. 



If, then, we compare the wave form of the current at the receiving 

 and sending ends when only a very short length of the cable inter- 

 venes, we find them practically the same. If, however, we test the 

 wave form at 20 or 30 miles distance we find it somewhat different, 

 and as we go farther and farther along the cable we see the wave 

 form not only become attenuated, that is, reduced in amplitude, but 

 smoothed out so that it has lost all the little irregularities of the 

 curve at the sending end. (See Fig. 2.) 



It is obvious that this is due to the harmonics of higher frequency 

 being more attenuated by the cable than the lower ones or funda- 

 mentals. 



This change in the wave form is called the distortion. If we 

 translate this fact into its equivalent in sound, it means that the 

 action of the cable is not merely to reduce the loudness of the sound, 

 but also to take away its quality, or those characteristics which give 

 it intelligible meaning, because these depend upon the presence of 

 the higher harmonics. 



A sound, therefore, at the sending end which has some distinctive 

 vowel-sound character becomes by the action of the cable so emas- 

 culated that at the receiving end it is recognizable as no vowel in 

 particular. 



This distortion is well shown by a series of photographs taken 

 with the oscillograph, showing the wave form at the beginning and 

 at the end of a cable of certain typical sounds. These oscillograms 

 were taken by Mr. B. S. Cohen in the Research Laboratory of the 

 General Post Office, and I am indebted to him for the loan of them. 



You will see how strikingly different the wave form of the 

 current coming out of the cable is from that entering it when certain 

 vowel sounds or syllables are uttered to a transmitter connected to 

 a receiver through 30 miles or so of standard cable. (See Fig. 3.) 



The problem presented was then to find the practical remedy for 

 this attenuation or decay of the amphtude of the waves, and especially 

 to cure the unequal attenuation of the short and long waves which, 

 according to Heaviside's theory, is the principal source of this dis- 

 tortion. 



It was long ago noticed that a leak or want of insulation in a 

 telephone cable will sometimes actually improve the clearness of 

 speech transmitted through it. Hence, one remedy for distortion 

 proposed by Professor Silvanus Thompson in 1891 was to put upon 

 the cable at intervals artificial leaks of a particular kind called 

 inductive shunts. It is, however, necessary to preserve a certain 

 loudness in the sound as well as clearness in order that speech may 

 be intelficrible, and it was the doubt whether this method when tried 



