WILLIAM SIEMENS, F.R.S. 51 



sistance C.' A scale containing resistance coils from 1 to 10,000 

 units would therefore only allow us to ascertain resistances not 

 exceeding these limits, but C and A being each composed of 

 o variable coils of 10, 100, and 1,000 units respectively, we are 

 enabled to measure any resistance between 0*01 and one million 

 units with the same degree of accuracy. By means of this arrange- 

 ment we measure the resistances of copper wire of any length and 

 the insulation resistance of long cables within the limits of correct- 

 ness of 0'2 per cent. 



For the insulation tests of short pieces of cables or of longer 

 cables of better insulating materials, such as india-rubber and 

 Wray's mixture, such method is no longer applicable, because 

 resistance coils of such diversity of dimensions as would be 

 necessary could not be used with sufficient accuracy, chiefly 

 because the greater battery power that would be required would 

 heat the smaller branches of the arrangement, and thus increasing 

 the resistance affect the result very considerably. 



It was therefore necessary to turn to another method for 

 ascertaining the value in units of the insulation resistances of 

 short pieces of cable, say one knot in length. We employ in such 

 cases a very sensitive sine galvanometer, or if the room permits 

 of it, a Weber's reflecting galvanometer of 40,000 turns, and a 

 magnetic reflector. 



By means of an adjusting magnet the sensibility of this instru- 

 ment can be varied between the limits of 1 and 100. 



The astatic condition of the needles of the sine galvanometer 

 being subject to changes, the constant of the instrument should be 

 verified repeatedly while testing. 



For the reading of this instrument in degrees we substitute 

 units of resistance by means of the following formula : 



(I.) 



in which R is the insulation resistance, < the angle of deflection, 

 <f> 1 the constant of the instrument, n the number of elements 

 employed. 



For the derivation of this formula, see Appendix No. 1. 



This method is applicable only for measuring great resistance 

 between certain narrow limits. During the progress of the cable 



E 2 



