386 
PROF. LOUIS VESSOT KING ON THE CONVECTION OF 
A/B = a/b = «//3 independently of any connecting- or contact-resistance in the bridge. 
It is this characteristic property of the Kelvin bridge which makes it especially 
applicable to the present series of measurements. 
In the actual apparatus the wire A to be tested is mounted in a specially 
constructed fork which is rotated at the extremity of a long revolving arm. The 
main circuit and the potential terminals PiPo are carried through four mercury- 
contact slip-rings to the overhead wires and thence to the bridge. 
In the actual apparatus the ratio coils had nearly equal resistances of the value 
10° ohms ; when compared by means of a Kelvin-Varley slide they were found to have 
the following ratios :— 
a. (3. a. b. 
24,984 24,986 25,020 25,010 
making the correction factor (a/b — a/(3) = 3/2500. 
The connecting resistance S was found to be about 0'55 ohm; the resistance B was 
never less than 0'50 ohm, so that the correction in equation (34) due to small 
departures from equality in the ratio-coils is given by 
A/B = a/b + S/B . 3/2500.(36) 
It will be seen that A=B to less than one-tenth of 1 per cent., which, as will be 
seen later, is well within the possible accuracy of the present experiment. We also 
notice from (35) that I does not differ from the ammeter current k by an appreciable 
amount. 
The galvanometer employed was a Broca instrument of resistance 97 ohms and 
adjusted to a sensitivity such that 8'5x 10 -10 amperes gave a deflection of 1 mm. on 
a scale at a metre distance—a shunt was used with the galvanometer so that the 
sensitivity could be reduced to any required fraction of this amount. 
The ammeter was a direct-reading Weston instrument, consisting of milli-voltmeter 
and shunt, by means of which ranges 0'2, 2'0, 20 amperes were available ; these ranges 
could be further subdivided by the insertion of a suitable resistance in series in the 
milli-voltmeter circuit. The instrument was calibrated against a Weston laboratory 
standard and was found to read correctly to within ^ of 1 per cent., which represents 
the order of accuracy aimed at through the experiment. 
The procedure carried out in order to obtain a set of readings was as follows : the 
wire under test was placed in the fork and potential terminals of fine platinum wire 
(about 1 mil diameter) fused on the wire at a measured distance apart. A 2-volt 
cell was inserted at TjT 2 and a Kelvin-Varley slide placed in parallel with the circuit 
SP^BA, S being a standard resistance connected in series with the circuit A, B so 
that the resistance of the wire in ohms might be obtained. By using a very small 
measuring current (about 0‘05 amperes) and the galvanometer at full sensitivity the 
resistance of the wire between potential terminals at standard room temperature 
