490 Dr. 0. H. Draper on the 



Preliminary Experiments. — A set of readings was first 

 taken with the temperature approximately constant at 18 0- 5 C, 

 and with rather weak currents, the tangent-galvanometer 

 having two rings and sixty turns of wire. The readings 

 were : — 



Electron gSSSt 



divisions deflection. 



26-5 -161 



27-3 -260 



28-6 -471 



29-5 -922 



30-0 ..... 1-340 



Here evidently the current was increasing faster than the 

 potential-difference ; and the inference was that the polariza- 

 tion was increasing in amount, and did not reach a steady 

 value with such currents as could be measured by this galva- 

 nometer. A galvanometer with a single ring was therefore 

 used in the greater part of the experiments. To obtain a 

 deflection as high as 45° on this galvanometer it was neces- 

 sary to use eight or ten Grove cells ; and the potential- 

 difference with such strong currents was such as to send the 

 spot of light off the scale when the distance from mirror to 

 scale was that for which the divisions represented equal diffe- 

 rences of potential. The scale was therefore moved nearer the 

 mirror and calibrated. A current was sent by five Leclanche 

 cells through 12,000 ohms resistance, while the electrometer 

 was joined to two movable points in this resistance. Between 

 these two movable points various resistances were included, 

 from 1000 to 12,000 ohms ; the readings being taken up one 

 side of the scale and back, then up the other side and back to 

 zero, in order to eliminate the effect of loss of charge in the 

 electrometer. The scale was thus divided into twenty-four 

 parts, corresponding to equal differences of potential, and a 

 table was drawn up from which the value of any deflection 

 could be obtained in terms of the central scale-division. The 

 scale was put as accurately as possible at the same distance 

 from the mirror on each day ; but in order to ensure absence 

 of error from displacement of the scale, as well as to have a 

 less arbitrary unit than a scale-division, the deflection pro- 

 duced by a standard cell was taken at the beginning and end 

 of each set of observations. 



The Standard Cell. — The cell chosen was a modified form 

 of the Daniell, resembling that described by Professor Lodge 

 in the Philosophical Magazine for January 1878. In a 



