614 PROF. J. J. THOMSON AND MR. G. F. C. SEARLE ON THE RATIO OF THE 
The box III. had one of its coils wrong wlien we began to use it, and during part 
of our work when we were attempting to use much greater battery power than we 
finally adopted, three more of the coils gave way. For this reason we could only use 
six out of the ten coils. These six coils, as well as box 11., kept their resistances 
very constant during the whole of the investigation. They were tested several 
times during a whole year, and no perceptible change was detected. 
The resistance boxes were placed on blocks of paraffin, and the insulation tested by 
a gold leaf electroscope. 
Method of making the Observations. 
The battery, resistance boxes, galvanometer, and condenser were connected as in 
the diagram (3), the commutator being placed so near to the condenser that a very 
short wire sufficed to make the connection. The insulation of the whole system when 
connected up was tested from time to time by a gold leaf electroscope. 
The electrically driven fork was set going and the beats of the auxiliary fork with 
the standard fork observed. The water supply was then adjusted so that the motor 
drove the commutator at the required speed. One observer (G. F. C. S.) observed 
the stroboscopic disc of the commutator through the slits of the fork, and kept the 
speed steady by means of the controlling arrangement already described. The other 
observer (J. J. T.) observed the galvanometer. When the speed of the commutator 
had got steady, resistances were taken out of the Wheatstone’s bridge box in the 
arm CD until no deflection was produced when the galvanometer circuit was broken. 
The sensitiveness of the galvanometer was such that the effect produced by altering 
the resistance in CD by 2 ohms in 3000 could be detected, and the speed of the 
commutator was kept so steady that the light reflected from the galvanometer 
miri’or did not move over more than half a division; the deflection produced by the 
condenser when not balanced was more than 500 scale divisions. 
The battery was then reversed and the operation repeated. The wire connecting 
the condenser to the commutator was then detached from the condenser, and the same 
operation repeated. In this way the capacity of the wire was determined. The 
temperatures of the fork and resistances were then read, and the beats of the auxiliary 
fork with the standard fork ao’ain determined. 
The results of these observations are exhibited in the annexed Table. 
