100 



Professor J. A. Fleming 



[March 27, 



or insulator, such as paper, guttapercha, etc., possesses some small 

 degree of true conductivity for alternating electric currents. This 

 conductivity, as shown by experiments made in my laboratory at 

 University College a year or two ago, increases with the frequency of 

 the oscillations and with the temperature. 



In the case of telephone cables, and especially loaded cables, the at- 

 tenuation depends, amongst other things, upon the ratio of the dielectric 

 conductivity (S) to the capacity (C) of the cable per unit of length. 



This ratio of S/C is rather large for guttapercha at ordinary 

 temperature and for a frequency of 800 to 1000. Hence the import- 

 ance of it. For guttapercha such as was used in the Anglo-French 

 cable, the value of S/C was about 120. Before making the Anglo- 

 Belgian cable, Messrs. Siemens discovered a method of preparing, or 

 treating it so as to reduce the value of S/C to 12 for that cable. 

 This is an enormous improvement on previous materials. The exact 



Tcf Tto«6«i|tteT 



lb Tic 



To Tni»va«l tti». 



s i 



Na2 Side Circuit 



To R«cc 



Fig. 7. — Arrangement op Side and Phantom Circuits for Phantom 

 Working in Duplex Circuits. 



nature of this mode of preparation has not been divulged. It may 

 be in part at least a most careful desiccation of the guttapercha, as 

 I have found that a very small amount of moisture has the effect of 

 greatly increasing this ratio for a dielectric. 



The Anglo-Belgian cable, like the Anglo-French cable, is a coil- 

 loaded cable, but is so constructed that by a simultaneous use of the 

 two circuits in it, it is possible to provide a third, or so-called 

 phantom circuit, obtained as follows : — 



The cable has in it two complete metallic circuits, each of which 

 can be used separately as an independent telephone circuit. The 

 circuits are connected at the two ends through induction coils, or 

 transformers, to their respective transmitters and receivers. If, 

 however, we use the two wires of one circuit as a lead, and the two 

 wires of the other as a return, we can make a third circuit, or 

 phantom circuit, and telephone along it quite independently of what 

 is going on along the other two main circuits. (See Fig. 7") 



