ELECTRIC DISCHARGES IN GASES AT LOW PRESSURES. 
421 
Observed Minimum Spark Potential in Volts. 
Spark length in 
millims. 
Air. 
Hydrogen. 
Carbon dioxide. 
1 
349 
278 
419 
2 
351 
279 
419 
3 
356 
278 
419 
5 
349 
281 
421 
10 
348 
283 
423 
wliej'e it will be seen that the values of the minimum spark potentials, for air, over 
this large range of spark lengths vary by only 7 volts. The values for hydrogen, 
over the same large range of spark leiigtlis, vary l)y only 5 volts, and those for 
carbon dioxide Ijy only 4 volts. 
These results, then, seem to establish the fact that the least sjmrk potential 
recphred to break down a gas is entirely independent of the s})ark length. 
It is evident, too, from hgs. 3, 5 and 7, that the constancy of the minimum spark 
potential is a necessary condition to Paschen’s law holding for discharges at 
diherent electrode distances. 
11. J. Strutt,^ in his ])aper “ On the Least Potential Differences recjuired to produce 
Discharge through Gases,” has drawn the conclusion that tlie minimum spark 
potential for discharges in any selected gas is j)robably equal to the cathode fall, in 
the same gas, measured over the whole extent of the negative glow in the vacuum 
tube. Since the cathode fall in any gas has been shown by Warburg! to be a 
constant, over a very large range of pressures, the constancy of the values obtained 
in these experiments for the least spark potential gives strong support to Strutt’s 
conclusion. 
Moreover, the value of the least spark potential found by Strutt in air, using a 
spark length of f millim., was 341 volts. This agrees very well with the numbers 
given above, the difference being only about 8 or 9 volts. For hydrogen, however, 
the agreement between the results is not so good, his values for the least spark 
potential, 302-308 volts, being somewhat higher tlian those found for hydrogen in 
these experiments. 
In this connection it may be pointed out tliat special precautions were taken in 
the neighbourhood of the critical pressure to make certain that the gas was in 
its normal condition when the discharge occurred, and so make sure that the 
s})ark potential obtained was not too small. The procedure followed was to apply a 
low voltage to the electrodes, and then gradually increase it until the discharge 
passed. By this procedure there could he no doubt that the gas was always in its 
* Strutt, ‘Phil. Trans.,’ A, vul. l'J3, p. 393. 
t Waururg, ‘Ainu d. Phys.,’ voL 31, p. 545. 
