339 



perature due to air currents as well as serving for yokes to the magnets. 

 A current of one-tenth ampere was used which insured a high state 

 of magnetization in the iron when two coils were in series, giving 5000 

 turns. 



The coils were connected in the bridge in such a way that the two 

 coils in one case formed the opposite arms of the bridge. By means 

 of a reversing switch the current in one of these coils could be reversed. 

 This changed the field which might affect two opposite arms of the 

 bridge and thus doubled the deflection. Another switch might have 

 been inserted in the other pair of arms and thus doubled the deflection 

 again but errors due to the switches would also have been doubled and 

 no advantage gained. The switch was carefully constructed with large 

 copper rods dipping into copper mercury cups but, at best, the inac- 

 curacies of the switch limited the accuracy of the experiment. 



The fine adjustments were made by resistance boxes shunted round 

 one of the coils. About 15,000 ohms in this shunt balanced the bridge. 

 A change of one ohm in the shunt gave a deflection of two millimeters 

 and indicated a change in the resistance of the arm of yinnnnr ohm. The 

 whole resistance being over 100 ohms this would give a determination 

 of one part in 2,000,000 or, since the deflection is doubled, one part in 

 4,000,000 for each arm. The result of 30 readings each way was that 

 the shunt resistance was about 3-4 ohms less with magnetic field than 

 without. The shunt was so placed that this gives a less resistance by 

 one part in 1,200,000 when producing a magnetic field. 



The above result is in the wrong direction. The difficulty may lie in 

 the fact that the galvanometer, though used at night, was unsteady at 

 best, or it may be due to leakage. The resistance of the coils was 100 

 ohms while the insulation resistance was 11,000,000 ohms. If the leak- 

 age is symmetrical along the doubled wire it will not affect the galvano- 

 meter upon reversing the current in one coil. This assumption may 

 not be justified. 



