226 Rey. P. J. Kirkby on the Passage of Electricity 
how initial conductivities through rarefied gases can be 
multiplied by increasing the electric force, owing to the 
formation of ions by the collision chiefly of the negative ions 
with the molecules of the gases. 
By an application of these methods, for any small pressure 
of the hydrogen and oxygen, the potential of AB could be 
brought as near the sparking potential as possible. More- 
over the passage of small sparks could be detected by the 
electrometer, even when they were invisible by daylight. 
The result of the first series of experiments on the effect of 
currents previous to sparking is a negative one. 
No perceptible combination was observed to take place 
without the passage of a spark, however near the potential 
of the lower plate was 1o the sparking potential. Thus 
previous to sparking, neither electric force nor ultra-violet 
light, nor the currents due to their joint action when ionization 
by collision takes place, had any measurable effect in the way 
of combining hydrogen and oxygen at small pressures. 
But directly a spark passed, a small measurable fall of 
pressure occurred. 
These results being established, the electrometer was dis- 
carded for the remaining experiments by which it was 
proposed, if possible, to determine the connexion between the 
fall of pressure and the current. The electrical connexions in 
the new arrangement are indicated by the diagram of fig. 3. 
so 2 
Fig. 3. 

Ais the bell-jar, U a resistance- box, Z a battery of about 1000 
volts, and Va voltmeter. The last was made on the principle 
of the D’Arsonval galvanometer, and so could be used equally 
well asan ammeter. It was specially constructed by Messrs. 
Nalder & Thompson with a high resistance (38,000 ohms), 
and it was graduated up to 200 volts. Thus currents of the 
order zt, ampere could be read accurately and quickly, for 
the instrument was dead-beat. Its reading was generally 
over 100, so that the error involved in determining the 
current passed through was a small one. 
