ELECTROMOTIVE FORCES OF THE ELECTRIC ARC. 315 



To form some idea of the magnitude of these capacity currents or leaks, the four 

 points a, b, c and the arc, were opened in pairs. One of the opened points was 

 re-connected through the thermo-galvanometer used as an ammeter to measure the 

 capacity current supplied to that part of the circuit between the instrument and the 

 other point that was opened. The direct current circuit was disconnected at the 

 points a and d. The alternating currents observed are given in Table I. for a "R.M.S. 

 P.D. of 3'65 volts and a frequency of 100,000 ^ per second. 



TABLE I. Capacity Currents in Leads. 



The maximum value of the capacity leak observed is 3 '9 X 1.0-' ampere at the 

 highest frequency used in any series of experiments, and at a P.I), about seven times 

 as high as that used, so that if the capacity current is proportional to the P.U., it 

 should not exceed O'G X lO" 4 ampere, or about O'OO per cent, of the working current 

 of O'l ampere, and may therefore be neglected. Even if this capacity leak had been 

 many times larger, it would not have appreciably affected the P.U.'s measured, since 

 at these high frequencies the arc behaves like a non-inductive resistance, and 

 therefore the measuring current and the capacity current would add approximately 

 as vector quantities at right angles. 



As it is only the relative values of V A , V K , and V that are required very accurately, 

 any small self-induction in the leads connected with the points 1, 2, and 3, or in the 

 measuring circuit itself, is of no importance. Nevertheless, all the leads were 

 carefully stranded together to prevent any E.M.F.'s being induced in this circuit, 

 caused by magnetic induction. The only wire which could not be stranded with a 

 corresponding wire was about 30 centims. of the lead between the fuse f 3 and the 

 movable contact t 3 . Experiments were made by varying its length and position to 

 see if it introduced any error, but none could be detected. 



The condensers F and G were placed some distance apart, so as to prevent any 

 direct electrostatic action between them. This, as well as any mutual induction, both 

 electrostatic or magnetic, between any part of the main or alternator circuits and the 

 measuring circuit was examined for, but none could be detected. Experiments were 



2 s 2 



