Electricity through Argon and Helium. 305 



connected together. The zero was read. The quadrants were 

 disconnected. The rajs were turned on for a convenient time, 

 usually twenty seconds. The number of divisions through 

 which the spot of light had moved was observed. 



The other gas cell was then substituted and the same 

 process gone through, and so on alternately. 



It was first necessary to make sure that the E.M.F. used 

 was sufficient to saturate the gas. For this purpose the rate 

 of leak through the air cell was determined with various 

 E.M.F.'s. 



The rays remained on 20 seconds in each case. 



No. of Cells. No ' of Prions passed 



over in 20 seconds. 



5 18 



10 27 



15 42 



20 55 



25 67 



50 78 



100 79 



Thus it was found that 100 cells scarcely produced a larger 

 current than 50. Saturation had been reached. In all the 

 subsequent experiments 100 cells were used. 



The two cells, both filled with air, were now compared. 

 The numbers given are the number of divisions passed over 

 in 10 seconds. 



Cell No. 1. Cell No. 2. 



53 T^^-62 



Another set of measurements gave the following results : — 

 Cell No. 1. Cell No. 2. 



50 Z^=^ 58 



The discrepancies between successive readings may be 

 almost entirely attributed to irregularities in the action of the 

 contact-breaker of the coil, and of the high vacuum tube. 

 The measurements following immediately upon one another 

 should be most strictly comparable. 



The best method of reducing these results is therefore to 



