a 
Genesis of Ions by the Motion of Positive Ions. 747 
irregular large deflexions took place at intervals when the 
tinfoil rings were insulated, or maintained at small potentials *. 
In order to determine the sparking potentials, a battery of 
small cells was connected to the electrodes through a volt- 
meter having a resistance of 38,000 ohms and an adjustable 
resistance. The number of cells was increased until a glow 
appeared in the gas and the voltmeter gave a deflexion when 
a beam of ultra-violet light acted on the negative electrode. 
The intensity of the light used in these experiments was very 
small, being a thirtieth of that used in the experiments for 
the determination of the values of « and B, so that the ions 
set free from the negative electrode would constitute a current 
of the order 3 .10-“ ampere. 
The values of the sparking potentials obtained by this 
method were very definite, and showed no irregularities 
depending on the length of time the gas is subjected to the 
electric force. The total external resistance which was used 
for the determinations which are given in the following tables 
was 48,000 ohms. It was found necessary to use a resistance 
of this amount so that too large a current should not pass 
through the gas when the sparking point is reached. Small 
variations occurred in the determinations when the external 
resistance or the intensity of the ultra-violet light was altered. 
When the intensity of the light was increased to thirty times 
the value which was used in the determinations, the potentials 
required to produce a spark were diminished by about three 
or four volts. When the resistance was increased by 20,000 
ohms, it required a few more volts to produce a spark in the 
gas. There would probably be no variation in the potential 
due to a change in the external resistance if an electrometer 
was used to detect a discharge through the gas, but other 
effects give rise to irregularities in this case. These differences 
between the determinations of the sparking potentials are 
very small, and do not affect the agreement with the theory. 
The following tables give the values of the sparking 
potentials V as found experimentally for the distance a 
between the electrodes. The numbers in the columns ax X 
are the values of the sparking potentials obtained theoretically ; 
a being the sparking distance for the force X and pressure p 
log a--log 6 
a—Bo 
* Unless some such precautions as these are observed the sparking 
potential for pressures greater than the critical pressure cannot be found 
accurately by a sensitive electrometer, and we think it probable that the 
“lao” which has been noticed by some observers is due to the effect we 
haye described, 
given by the formula a= 
