42 Mr. 0. Heaviside on Duplex Telegraphy. 



exact balance being always required, the signalling can be 

 continued for great lengths of time without any change of ad- 

 justment ; and, moreover, the balancing resistances may some- 

 times be altered very considerably without actually interrupt- 

 ing the signalling. The actual received current may be con- 

 sidered as the algebraical sum of two parts — one the proper 

 received current, the oilier an interfering current produced by 

 inexact balance. In the double-current Morse system in com- 

 mon use in England the marks are made by one current (say, 

 positive), and the spaces by the negative current. If C is tho 

 strength of the received current, then the whole range of the 

 current is 20. In the single-current Morse system employed 

 on the Continent and elsewhere there is no current during the 

 spaces ; hence the range of the current in the receiving instru- 

 ment is only C. Therefore an instrument that admits of being 

 worked either by single or double currents, as magnetized or 

 polarized relays, will give signals twice as strong with double 

 currents as with single with the same battery-power. This 

 applies both to ordinary single working and to duplex work- 

 ing. In the latter there is a further advantage in favour of 

 double currents. It is theoretically possible to work duplex 

 with double currents when the interfering currents are little 

 less strong than the received currents ; for as the received 

 current is always either + or -— C, the superposition of any 

 current of less strength than C will not alter the sign of C, 

 whether + or — . On the other hand, in single-current work- 

 ing the received current is always either C or zero. In the 

 first case the current C overpowers the tension of a spring or 

 other opposing force ; and in the latter the spring is unopposed. 

 The most rapid signalling is to be obtained when the forces 

 moving the armature or tongue of the relay are equal in each 

 direction ; and then the retractile force of the spring must be 

 equivalent to a reverse current of the strength ^0. Therefore 

 the interfering currents in duplex working with single currents 

 must never be so great as JC — thus giving an immense advan- 

 tage to the double-current system as regards freedom from in- 

 terruption by inexact balance or other causes, in addition to 

 the advantage before mentioned of giving signals of twice the 

 strength. It is found by experiment that duplex working 

 (Morse) will not be actually interrupted until the interfering 

 currents are as much as \ or \ the strength of the received 

 currents with double currents, and \ or \ with single cur- 

 rents — although no hard and fast line can be drawn, owing to 

 the very numerous causes in operation. On an overland wire 

 worked duplex with differential relays and double currents the 

 resistance which gave exact balance was, at one end, 2560 



