178 ROYAL SOCIETY OF CANADA 
distances, 1, 2, 3, 5, and 10 mm. The carbon dioxide was prepared 
by treating marble with hydrochloric acid and was purified and dried 
by being bubbled through a wash-bottle of water and passed through 
a tube tightly packed with phosphoric pentoxide before reaching the 
discharge apparatus. In each case the operation of exhausting the 
whole discharge apparatus to 1 mm. or less, of mercury, and then 
refilling with carbon dioxide, was repeated five or six times, and finally 
the gas was allowed to stand, as in both previous cases, in the presence 
of a bulb of phosphorus pentoxide for several hours. 
The complete set of results is given in Table V. and the corres- 
ponding curves set forth in Fig. VI., and if we again compare the 
discharge pressures and spark lengths corresponding to any value of 
the applied potential, the same law is seen to hold here also with even 
greater rigidity than in the other cases. 
For 1800 volts the figures are approximately: — 



Distance between electrodes Discharge pressures in 
in mm. mm. of mercury. 
1 | *817 
2 421 
3 ‘274 
5 ‘164 
10 "0892 




where the pressures are almost in the required ratio, being 9-2: 4:8: 
DUO 
For 500 volts the numbers are :— 



Distance between electrodes Discharge pressures in 
in mm. mm. of mercury. 
2°34 
2 1°23 
3 ‘84 
5 7 
10 "28 


where the pressures are as 8-4: 4:4: 3: 2:1. 
