A. E. Kennelly — Alternating Currents. 457 



rent c in absolute measure equivalent to the passage of this 

 quantity Q, we have 



c = n n e h X lO -7 (2) 



and Y = 5*43 n e h t X 10 -7 (3) 



Besides the Cardew voltmeter steady potential indication,^ 

 quadrant electrometer, not shown in the diagram, with its 

 needle connected to one pair of quadrants, was employed to 

 measure the potential difference at the voltameter terminals 

 and thus indicate its mean resistance. 



The preceding table gives the results of 10 measurements. 

 The volume of the gas is corrected for temperature, but no 

 correction is introduced for the barometric pressure or the ten- 

 sion sustained by the gas in the tube. 



The condenser used was adjustable from 0*001 up to 5 mi- 

 crofarads, and from absolute determinations is probably accu- 

 rate to at least 0'5 per cent. 



The last two columns show that 



When the density of the percentage of gas actually 



current per sq. cm. of generated of that calculated 



either electrode was : as due was : 



1-12 6-2 per cent. 



1-9 38 



30-4 (average) . 82*6 " 



31-5 78*4 " 



46-1 (average) 89-3 



50-1 .._ 100-8 " 



62-6 (average) _._. 95*3 " 



69-7 .'. *.. 98*0 " 



141-8 95-6 



The trials Nos. 13, 14 and 15 were taken without condenser, 

 and with non-inductant resistances in its place. In trial No. 

 13 this resistance was a length of fine platinum wire offering 

 720 ohms (hot). In trial $Tos. 14 and 15 the resistance was a 

 96 volt 10 candle-power Edison lamp offering 230 ohms resist- 

 ance at equivalent illumination. In calculating the current 

 strengths for these three observations, the resistance of the 

 voltameter itself was required. It was measured in two ways : 

 1st, by means of the quadrant electrometer and the potential 

 differences it indicated on the known resistance and on the 

 voltameter ; 2d, by rapidly substituting a non inductant resist- 

 ance alternately with the voltameter in the lamp circuit and 

 adjusting this resistance until the illumination was equal in 

 both cases. These two methods concurred in showing the 

 resistance of the voltameter to be 30 ohms, and the maximum 

 resistance it offered through the whole series was 35 ohms. 



The table shows that 



