.s/y? WILLIAM SIEMENS, F.R.S. 79 



not exceeding, practically, one in one thousand. It was desirable, 

 sometimes, to determine fractions of units in measuring copper 

 conductors ; and at others millions of units in measuring the gutta- 

 percha resistance of a short piece of cable, to accomplish which 

 tip apparatus could be modified in different ways. Another 

 feature of this method of testing consisted in the close observance 

 of the time during which the electric current was allowed to act 

 before the observation was taken. This was of the utmost 

 importance, in order to obtain results that could be relied upon, 

 for the conductivity of gutta-percha was changed, even for days, 

 by the application of electric currents. Their method of ascer- 

 taining the inductive capacity of cables was also peculiar, being 

 based upon their discovery that inductive tension, in passing from 

 the conductor through the insulating covering, followed the simple 

 law of Ohm regarding electric condition, and admitted, therefore, 

 of being subjected to the same precise methods of measurement. 

 Although this system of testing cables had not been long in use, 

 resistance coils had been employed by them since the year 1849 for 

 determining the position of faults in subterranean lines. 



In the case of the Rangoon cable each mile of core was tested 

 after submersion during twenty-four hours in water at a temperature 

 of 75. Comparative testing would be useless unless made at the 

 same temperature, because the conductivity of gutta-percha 

 increased in very unequal ratio with the increase of temperature. 

 After submersion the cable was placed in Reid's pressure tank in 

 order to discover the existence of any cavities in the covering, but 

 the pressure that could be applied was insufficient to force the 

 water into the cavities of the lower coatings. The results of the 

 electrical tests were then noted in tables reduced to units of 

 resistance per nautical mile. Having thus obtained a complete 

 record of the copper and gutta-percha resistances of each mile of 

 cable, it was sent to the wire-works to be covered with hemp and 

 iron. By this complete record or table it would be possible to 

 detect the slightest fault, where lengths of the core equal to, say 

 one hundred miles, had been joined together ; the copper resistance 

 should not, in that case, exceed the sum of all the resistances 

 contained in the table, due allowance being made for change of 

 temperature, whilst the resistance of the gutta-percha should not 

 be less than the sum of the resistances divided by 100. If it 



