1903.] 



Electric Discharges in Gases at Low Pressures. 



375 



discharges cat low pressures in all probability did not take place along 

 the shortest path between the plates, and it is inferred that the failure 

 of his numbers to establish the applicability of Paschen's law at all 

 pressures is clue to his having always taken this shortest distance 

 between the electrodes as a measure of the spark length. 



In the present paper an account is given of an investigation on the 

 potential difference necessary to produce discharges in a gas with a form 

 of apparatus which ensured the passage of the discharge in a uniform 

 electric field at all pressures. With this apparatus the spark potentials 

 were determined in air, hydrogen, and carbon dioxide, for different 

 distances between the electrodes, over a range extending considerably 

 above and below the critical pressures. Electrodes of brass, iron, zinc, 

 and aluminium, of the same size, were in turn used in the apparatus, 

 but the readings obtained showed that the spark potentials were not 

 influenced at any pressure by the size of the electrodes, provided the 

 discharge took place in a uniform field. 



The result of the investigation not only confirmed the truth of the 

 law enunciated by Paschen for discharges in a gas at high pressures, 

 but also demonstrated the applicability of the same law to the critical 

 and to lower pressures. This law is summarised in the statement 

 " that with a given applied potential difference, electric discharge in 

 a uniform field in any gas is dependent solely on the constancy of 

 the quantity of matter per unit cross-section between the electrodes." 



It is shown that a general application of Paschen's law demands 

 that the minimum spark potential must be a physical constant for 

 each gas. A special set of observations gave the following values of 

 this quantity for a number of simple and compound gases. 



Minimum spark 

 Gas. potentials in volts. 



H 2 278 



2 455 



H 2 S 414 



C0 2 419 



N 2 420 



S0 2 ; 457 



C 2 H 2 467 



Adopting Strutt's value of 251 volts for nitrogen, the conclusion is 

 drawn that the minimum spark potential is a property of the atom 

 rather than the molecule of a gas, and it is shown that if H', N', 0', 

 etc., represent the spark potential constants in volts, corresponding to 

 atoms of the gases H 2 , N 2 , 2 , etc., respectively, the minimum spark 

 potential for any compound gas whose formula is ELcN^O^, etc., will 

 be given by xS! + yW + zO' + etc. volts. It is pointed out that oxygen 

 forms an exception to the general application of this law. 



