1886.] electric discharge in a uniform electric field. 405 



considerable extent over the surface. Then in the gas between 

 the electrodes we have, when the discharge has just commenced 

 next to the negative electrode, a region where the electric field 

 has been discharged, and beyond this a region where the molecules 

 have not been dissociated but into which a few atoms from the 

 dissociated molecules have been projected by the explosion con- 

 sequent upon the electrical discharge. Now in the region next 

 to the negative electrode the molecules are in a polarized state, 

 since those molecules which were moving in a definite direction 

 have been dissociated. We may expect this polarization of the 

 molecules to produce an electric field, the effect of which on the 

 field, away from the negative electrode, is of an opposite character 

 to that of the original field which produced the electric discharge ; 

 if it produced a field of the same kind in this region there would 

 be instability. 



Thus the field outside this layer will be less intense than 

 the original field inside it, or than it was outside before the layer 

 of gas next the negative electrode broke down, so that if it did 

 not break down at first, it is not likely to do so afterwards. This 

 would evidently have a tendency to confine the discharge to the 

 neighbourhood of the negative electrode, which was a prominent 

 feature in our experiments with air and coal gas. 



Actual measurements by Hittorf and others of the electric 

 force in a discharge tube of the ordinary kind have shewn that 

 this force is very much less outside the first luminous patch which 

 we meet with as we travel from the negative electrode, than it is 

 between this bright patch and that electrode, and this being so it 

 is easy to understand why the discharge should be greatest near 

 the negative electrode. 



We shall now go on to apply the views we have been dis- 

 cussing to the case of the electrical discharge through a gas at a 

 low pressure, say less than 2 mm. of mercury. Before doing so 

 however it may be convenient to describe the phenomena observed 

 when the discharge passes through a tube filled with such a gas. 

 Suppose we start from the negative electrode ; in contact with 

 this we have a bright glow whose spectrum does not bear much 

 relation to the gas in the tube, it is presumably due to gas 

 absorbed by the electrode or by glowing matter from the elec- 

 trode itself. Next to this we have a dark space whose dimensions 

 depend upon the density of the gas, getting larger as the density 

 diminishes ; next to this we have a luminous mass of gas whose 

 spectrum shews bright lines and whose dimensions depend upon 

 the shape and &ize of the electrodes and the density of the gas ; 

 next to this we have a non-luminous space, for which the name 

 dark interval has been proposed by Dr Schuster ; after this we 

 have a glow often striated, reaching up to and touching the positive 



