July 3, 1884] 



NA TURE 



231 



trode to be of the same potential. The lines of force will then 

 cut the surface at right angles, and could we assume the con- 

 denser to be infinitely thin, there would only be a normal force 

 acting on its particles ; but as the lines of force are curved, the 

 particles not in immediate contact with the surface are acted on 

 by a tangential force which will tend to drive them away from 

 the positive electrode. As a steady state will only be possible 

 when the total force is normal throughout the condenser, we 

 arrive at the condition for the steady state that within the con- 

 denser the fall of potential must be the same for equal distances 

 measured along the normal to the surface. 



Experimental evidence speaks strongly in favour of such a 

 conclusion. If, for instance, a thin wire is used as electrode, it is 

 well known that the tension at the end of the wire before dis- 

 charge is very much larger than anywhere else. At high pres- 

 sures the discharge passes indeed from the end of the wire, but 

 as the exhaustion proceeds, the glow gradually covers the whole 

 wire, and the same amount of electricity flows out of equal areas 

 situated anywhere on the wire, for the dark space which alters 

 its width with the intensity of current is everywhere the same ; 

 this implies that the fall of potential per unit distance is the same 

 all over the wire. 



Hitherto we have only assumed a certain number of particles 

 positively electrified in the immediate neighbourhood of the 

 negative electrode, and we have left it altogether undecided 

 what these particles are. But if we consider now the fact that 

 the glow does not appear opposite the positive electrode, that is 

 to say, that while the fall of potential is the same all over the 

 surface the flow is stronger at some places than at others, we 

 are driven to the conclusion that the flow does not altogether 

 depend on the fall of potential, and we must again look for an 

 explanation in the chemical as well as the electric forces. 

 Wherever the fall of potential is chiefly produced by the presence 

 of the positively electrified particles, which I now assume to be 

 the decomposed molecules of the gas, these will help by their 

 chemical action to decompose other molecules. Opposite the 

 positive pole the fall of potential is principally due to near- 

 ness of that electrode ; chemical forces are absent, and the 

 molecules will not be decomposed. This is, I believe, the 

 explanation of the dark area. And it brings with it the expla- 

 nation of a large quantity of other facts, as, for instance, the one 

 which has been so long observed and well established, that once 

 a current is set up in the gas it requires a much smaller electro- 

 motive force to keep it going. For the discharge, according to 

 Us, will generally be introduced by a spark which must give the 

 first supply of decomposed molecules before the continuous glow 

 discharge can establish itself. 



I may for the sake of clearness once more mention shortly the 

 principal points of the argument. 



The rapid fall of potential in the neighbourhood of the 

 negative electrode renders the presence of positively electrified 

 particles in its neighbourhood necessary. 



If the distance through which the condenser action takes 

 place is sensible, the positively electrified particles will be acted 

 upon by a neighbouring positive electrode. 



A steady state will be established in which the fall of 

 potential along the normal from the surface will be everywhere 

 the same. 



As however the flow is stronger away from the positive elec- 

 trode, we must conclude that other forces besides electrical 

 forces determine the flow. 



It is natural to assume that these are chemical forces : that, in 

 other words, the positively electrified particles are the decom- 

 posed molecules, which by their presence assist the decomposi- 

 tion of others, and therefore the formation of the current. 



Unless a flaw is detected in this line of argument, I think 

 that the conclusion must be granted, namely, that the decom- 

 position of the molecules at the negative electrode is essential to 

 the formation of the glow discharge. This is really all that I 

 endeavour to support in this paper. The rest can only be 

 settled by further experiments. And amongst the rest I count 

 also the primary cause which originally produces the decompo- 

 sition of molecules at one pole rather than at another. It is 

 possibly due to an electromotive force of contact between the 

 gas and the electrodes which tends to make the gas electro- 

 negative. 



The gaseous molecules, then, according to our theory, are de- 

 composed at the negative pole. Their negative constituents can 

 follow the electric action, and as the fall of potential in the im- 



mediate neighbourhood of the pole is very rapid, the atoms will 

 I leave the pole with considerable velocity. That the region of 

 the dark space is filled with matter projected from the negative- 

 pole follows almost conclusively from the experiments of Gold- 

 stein and Crookes, and is also shown in a most striking way by 

 an experiment due to Hittorf. If a tube contains two parallel 

 1 wire electrodes at a distance of say a quarter of an inch, the 

 discharge will al high pressure pass in the usual way from elec- 

 ■ trode to electrode, but at very low pressures the discharge from 

 the positive pole goes away from the negative. The results can 

 be shortly expressed by saying that, as far as the positive pole is 

 concerned, the inner boundary of the dark space forms the 

 negative electrode. If the dark space is small and does not 

 reach to the positive pole, the discharge passes from the latter 

 towards the negative pole, but as soon as the dark space extends 

 beyond the positive pole, the positive part of the discharge goes 

 towards the nearest point of the dark space that is straight away 

 from the negative pole. 



We have then two closely adjoining, almost overlapping parts, 

 in which the discharge is in opposite directions, and this could 

 not be unless electricity is carried by matter which can, owing 

 to its inertia and high velocity, move against the electric forces. 

 To my mind this experiment proves conclusively that the nega- 

 tive electricity is bound to matter projected with high velocity 

 away from the negative pole. 



Goldstein has shown that when a thin pencil of the negative 

 glow belonging to one electrode passes close to another the 

 pencil is deflected. According to our view, such a pencil would be 

 formed by a succession of negatively charged particles pro- 

 jected in nearly the same direction away from the negative 

 electrode ; as these particles pass by another kathode, they 

 are naturally deflected out of their path by the electric forces. 

 Goldstein has shown that if the current is equally divided between 

 the two kathodes, the deflection is independent of the intensity 

 of the current, the pressure, and the nature of the gas. This 

 is exactly what ought to happen according to our theory, for 

 strengthening the current at one kathode means, as will presently 

 appear, increasing the velocity of the particles. The square of 

 the velocity will increase in the same ratio as the total fall of 

 potential in the neighbourhood of the negative pole ; as the 

 particles pass the other kathode, the forces from it are increased 

 in the same ratio as the square of the velocity with which they are 

 moving, and consequently the path will remain the same. Simi- 

 larly all the other experimental facts established by Goldstein 

 can be easily explained. 



The most conclusive proof of the view adopted in this paper 

 would be found in the demonstration that the amount of elec- 

 tricity carried by each particle was always the same, whatever 

 the current. I propose to test this fact in the following way : — 

 It was found by Hittorf that the particles proceeding from the 

 negative electrode, and projected at right angles to the lines of 

 force in a magnetic field are bent round in a circle. This is as 

 it should be, and I calculate that the radius of the circle ought to- 

 vary as s/F/e, where F is the total fall of potential within the 

 region in which the particles acquire their velocity, and e i- 

 the amount of electricity carried by each particle. As the 

 current increases, it is shown by Hittorf that F increases ; and 

 I find that at the same time the diameter of the ring in the maj' 

 netic field increases. If this diameter varies as the square root 

 of F, it would be proved that e must be constant as it is in elec- 

 trolysis. At present we can only say that the average amount 

 of electricity carried by the particles must increase less rapidly 

 than the fall of potential. Iff varies at all, we should expect it 

 to vary proportionally to the fall of potential in the neighbour- 

 hood of the negative electrode, and in that case the diameter of 

 the ring would be independent of the current, which it is not. 



The theory which I advocate involves the existence of a polar- 

 isation, and it might be considered a difficulty that no polarisation, 

 currents have with certainly been observed in gases. I believe 

 the difficulty only to be apparent, for the experiments prove that 

 the fall of potential near the negative pole, though rapid, is not 

 sudden, so that the layer within which the condenser action, 

 takes place is very much thicker in gases than in liquids. The 

 capacity of the condenser is therefore smaller, and though the 

 total fall of potential in the gas may even be stronger than in 

 the liquid, the polarisation currents might escape observation. 



With regard to the positive part of the discharge it will be 

 sufficient here to mention that stratifications are principally ob- 

 served in mixtures of gases or in compound gases, and that ii 



