170 ROYAL SOCIETY OF CANADA 
where the pressures are in the ratio 1:00: -55: -34: -22: -11, numbers 
which are again very nearly inversely proportional to the distance 
between the electrodes. 
Further, we notice that the spark potential corresponding to the 
critical pressure in all cases was practically the-same, 350 volts, and 
the values of the critical pressures for the different spark lengths 
were, from Table L.:— 


Distance between electrodes Discharge pressures in 
in mm. mm. of mercury. 
1 4°98 
2 2°71 
3 1°89 
‘5 1°34 
10 “679 


and ‘these numbers while not exactly in the ratio 10: 5: 3: 2: 1, are 
still very close to it. 
In finding the values for portions of the curves around the critical 
pressures the results given in Table I. show that a small variation in 
potential difference was associated with a relatively very large change 
in the pressures, so that a very small error in reading the potential 
difference would result in a large error in the pressure readings. It 
is interesting to note, however, that even under these unfavourable 
conditions a striking agreement is presented between the results 
obtained at critical pressures and the results demanded by Paschen’s 
law. 
In order to make the agreement between the numbers demanded 
by Paschen’s law and those obtained in these experiments still more 
evident, the results recorded in Table I. are again given in a slightly 
different form in Table II, where each potential difference is asso- 
ciated with the product of the pressure at which discharge took place 
and the corresponding spark length. Paschen? found that at high 
pressures these products were constant for different distances between 
the electrodes, as long as the applied potential difference was the 
same. 
The numbers recorded in Table II. show that the same law is 
rigidly applicable to all pressures both high and low. 
1 Paschen, Ann. d. Phys., Vol. 37, p. 69. 
