﻿930 Mr. A. H. Davis on Natural 



power of the values given. Also, as only moderate heating 

 of the wires was involved, little account has been taken of 

 temperature changes of the physical properties of the liquids, 

 though the point is referred to later. 



The ratio arms were usually 100 ohms each, 1000 ohms 

 being occasionally used for the larger currents. Equality of 

 the ratio arms was tested by reversing the connexions and 

 noticing whether the balance was thereby disturbed. 



To eliminate disturbing resistance changes when the heating- 

 current was passed, all connecting wires v^ere compensated 

 by similar wires in the opposite arms of the bridge, and the 

 resistances It and <r/R were constructed of manganin as also 

 was the slide wire S. It was found on test, by removing 

 the platinum wires TT and CC and substituting heavy oil- 

 immersed manganin of the same resistance, that the balance 

 of the bridge obtained with a small current CO 01 amp.) was 

 undisturbed when a heavy current of 1 amp was passed, and 

 that therefore the compensation was satisfactory. A 

 terminal head on the heavy copper leads to the wires 

 facilitated short circuiting of the platinum wires when 

 subsequent occasional test of the apparatus was required. 



It was possible that the heating current might set up 

 thermoelectric effects, but a test showed that these were 

 absent. The method was to switcli off the heating current 

 after it had been adjusted to the value necessary to balance 

 the bridge, and no appreciable drift of the galvanometer was 

 then observed while the wires were cooling- 

 Thermoelectric effects might also have existed through 

 local inequalities of temperature caused otherwise than by 

 the heating current, and these would reveal themselves by 

 drift of the galvanometer on making the galvanometer 

 circuit while the battery circuit was still open. Some slight 

 effect was occasionally found, but it was only important 

 when getting the original balance with the small current ; 

 and for this it was eliminated by the familiar device of 

 working with the galvanometer circuit always completed 

 and adjusting the bridge for no deflexion when the battery 

 was reversed. 



A large storage battery was used at first, and it gave 

 troublesome galvanometer effects due to leakage, discovered 

 and traced by reversal of the battery and galvanometer 

 circuits. To eliminate these a separate 10-volt battery was 

 used, and this and all parts of the apparatus and leads were 

 insulated by standing upon ebonite disks, etc. 



The increment resistance <£R was obtained from a box 

 of manganin coils (resistance 0*01 to 1*0 ohm) provided 



