SUBMARINE TELEGRAPHY 225 



be treated as one conductor, of which the observer has two 

 ends in his possession. He can then arrange his test so that 

 his observations at both ends are really simultaneous, with the 

 fault in the same condition when added to the two circuits. In 

 this case, a test based on the above principle is quite perfect, 

 and will fix the position of a fault with great nicety. But 

 where the cable has only one conductor, the two tests must be 

 made by different observers at different times. Faults have a 

 disagreeable art of varying very rapidly, so that their resistance 

 is never the same for two minutes or fractions of a minute, and 

 then the test becomes inaccurate, though not actually useless. 

 For instance, the observer in the first case might feel quite 

 sure that the fault was not more than ten miles off, even if he 

 got no information from the other end ; if the fault were caused 

 by a nail joining the copper and iron of the cable, it would 

 have no sensible resistance, and the above test would show it 

 was exactly ten miles off. Even if the cable were broken, 

 the observer could guess from the variation of the fault, the 

 current it returned, and other peculiarities, whether it was 

 likely that the fault had much resistance, and thus form by 

 the aid of experience a fair guess at its exact position. 



The measurement of resistance is far from being the only test 

 of which the results can be expressed with numerical accuracy ; 

 for instance, the statical tension at any point of the wire, its 

 potential, as it is called, can be measured by electrometers, and 

 indirectly by various methods. This statical tension is the quality, 

 in virtue .of which one electrified body attracts or repels another 

 more or less strongly. When a current is flowing from a battery 

 through a conductor to earth, the potential gradually decreases 

 from a maximum at the battery to zero at the earth, and de- 

 creases according to well-known laws. The observation of this 

 potential at any point gives additional information, therefore, 

 by which the condition of the conductors may be determined. To 

 revert to the analogy of the water-pipe, the potential would be 

 represented by the pressure per square inch, or head, inside the 

 pipe at each point ; it would be greatest near the cistern, and 

 gradually decrease to nothing at the mouth of the pipe where 

 the water was discharged. 



Another class of fault is more easy to manage. If by acci- 



VOL. II. Q 



