Passage of Electricity through Gases. 167 



movements that it appears almost continuous. We then have 

 a real galvanic current, upon which the magnet acts as upon a 

 rigid conductor, and which is capable of effecting electrolytic 

 decomposition of the gas, if it is compound, as in a liquid 

 electrolyte. In this case the resistance will be in proportion 

 to the length of the column of gas. It is therefore no longer 

 a discharge, but a real electric current." In proof of this 

 view it is especially urged that when induction-currents are 

 employed, besides the disruptive discharge, the "aureole" 

 appears, as well as the " trait de feu," which disappears in 

 highly rarefied gases, and, according to Perrot's experiments, 

 these can be separated from one another. 



But if an extremily rarefied gas acted in the same way as a 

 metallic conductor or an electrolyte, the smallest difference of 

 tension at the electrodes would cause a proportionate current 

 to pass through the gas, and the electricity could not, as ex- 

 perience shows, accumulate before the commencement of the 

 discharge in a greater but finite density. 



To make it possible to combine the latter phenomenon with 

 the supposition of an electrolytic conduction of the gas, we 

 must form the hypothesis of a certain resistance at the elec- 

 trode itself, which, like a non-conductor, would prevent the 

 passage of electricity from the electrodes to the conducting 

 gas, and to break through which a finite tension would be 

 requisite. 



The resistance itself could only have its origin in the 

 stratum of gas condensed on the surface of the electrodes, 

 the density of which is partly due to external pressure. 

 But the experiments upon galvanic polarization by gas at 

 diminishing pressures show that the density of a gas stratum 

 condensed upon metal is very little influenced by the pressure ; 

 while on the other hand the experiments upon the electric 

 discharge show that the mass of electricity necessary for the 

 production of a discharge, and therefore the supposed re- 

 sistance, increase with the pressure. If the electrodes be 

 heated to about the temperature of melting glass, the gas 

 stratum should also change ; but the quantity of electricity 

 remains remarkably constant if the pressure be unchanged. 

 Furthermore, it would follow from these experiments that the 

 resistance is greater at the positive than at the negative 

 electrode, while yet we cannot perceive why greater power 

 should be required at the negative than at the positive to 

 break through the stratum of condensed gas. To explain these 

 different conditions at the two electrodes, we should have, be- 

 side the supposition of resistance, to form a second hypothesis, 

 somewhat like that mentioned in the former memoir, that the 



