100 Major Cardew. On obtaining a Unidirectional [May 31 r 



Fia. l. 



sensitive D'Arsonval reflecting galvanometer, S its | shunt, Ii and I 2 

 impedance coils, calculated to pass a current of less than 0*005 

 ampere with the whole alternating pressure in use on the system 

 between the terminals, L a non-inductive resistance formed of four 

 50 C.P. 50-volt. incandescent lamps in parallel, E a connexion to the 

 iron water-pipes supplying the station. 



The object sought to be attained by the use of this arrangement 

 was to obtain an indication of any leakage on the alternating system 

 by a method which would be unaffected by the capacity effect of a 

 large system. 



It is intended to substitute for the D'Arsonval galvanometer used 1 

 in these tests, a form of siphon recorder, so as to obtain a continuous 

 record of leakage. 



1 In the first tests, made on the 25th April, 1894, the mains in 

 connexion consisted of eleven circuits all connected to one machine. 



The pressure in the alternating circuit was rather greater than 

 1000 volts, and about half this pressure was indicated by an electro- 

 static voltmeter between M 2 and earth throughout the experiments. 



The battery used was six cells, and the following deflections were 

 obtained. 



With + ve pole of battery to the mains . 

 With battery out of circuit 



20 to left, 

 140 to right. 



48 



Various modiBcations were tried, but in all cases the results showed 

 an apparent electromotive force of from 5 to 6 volts, tending to cause 

 a flow of positive electricity to the water-pipe earth. 



In order to settle this question, a small copper voltameter, consist- 

 ing of two No. 40 S.W.G. copper wires in CuS0 4 solution, was 

 inserted in place of the galvanometer and the shunts were removed. 



