ELECTRIC DISCHARGES IN GASES AT LOW PRESSURES. 
431 
and if we use these values to calculate the minimum spark potentials in the remain¬ 
ing gases, Ave obtain : 
Gas. 
Value found l)y 
experiment. 
Calculated 
value. 
C-IL. 
468 
474 
N..0 
418 
412 
oh 
455 
321 
The ao'reement between the observed and calculated wilues for eacii of tlie u’ases 
N,jO and CnHo is very marked, and is a strong evidence that the additive law holds. 
The only case in which there is any serious difterence lietween the observed and the 
calculated values is that of oxygen. Judging tliat this discrepancy might lie due to 
impurities, three s])ecimens of this gas Avere prepared by independent methods. It 
Avas prepared in turn by electrolysis, by heating potassium permanganate, and by 
heating a mixture of potassium chlorate and manganese dioxide. In every case the 
gas Avas |)urihed by being passed rlirough a mixed concentrated solution of 
potassium iodide and caustic potash and through concentrated sulpliuric acid. It 
Avas also carefully dried in the usual AA-ay. It will be seen from Table VII. that the 
three sets of readings jiractically coincide at every pressure, and, since it is not 
possible that the same impurity could be present in each of these specimens to the 
same degi'ee, it does not seem reasonable that the Irregularity in oxygen could be 
traced to impurities arising from any lack of precaution in the preparation of the 
gases. 
It is Avell known, however, tliat when an electric discliarge Is passed through 
oxygen, a considerable tpiantity of ozone is produced. It is in fact by this method 
that ozone in its purest form can be obtained. It is highly probable, then, tliat 
after the first discharge had passed betAA^een the electrodes, in tlie experiments on 
oxygen, a considerable percentage of ozone was present in the gas, and it may lie that 
the discrepancy noted abiove is due to this cause. The experimental Audue of 
455 volts found for oxygen .seems to bear out tills conclusion, for, ascribing the 
value of 161 volts to the atom of oxygen, we get by addition 483 Amlts as the calcu¬ 
lated Amine of the minimum spark potential for ozone. 4Iie difference between the 
tAvo values is but 28 volts, and assuming that the di.scharge occurs Initially through 
the dissociation of ozone rather than of oxygen, the result Is not in opposition to the 
additive laAv Av-hich has been shoAvn to hold for the other gases. This large influence 
of a small amount of a denser gas Avhen mixed aaIjIi one less dense is In accord with 
the results obtained by previous expeiimenters, for Wauburg'^' and CapstickI in 
their experiments on the cathode fall, and Strutt| in his experiments on the 
* Warburg, ‘ Wied. Ann.,’ vol. 31, p. .545. 
t Cafstick, ‘Roy. Soc. Proc,,’ vol. 63, p. 360. 
j Strutt, ‘Phil. Trans.,’ A, vol. 103, p. 385. 
