496 



NA TURE 



[March 22, 1906 



conditions of the tube makes the molecules pass from a 

 condition in which they are not giving out an appreciable 

 amount of light to one where they are brightly luminous, 

 and, as the great increase of the current when the 

 luminosity appears shows, this change in state is accom- 

 panied by an emission of corpuscles. From this and other 

 considerations ,1 have come to the conclusion that what 

 takes place when the gas becomes luminous is that the 



Discharge 



Discharge 



10 FlFLO eo 



internal energy in the atom, in consequence of its 

 bombardment by the corpuscles, increases, and when it 

 gets up to a certain critical value the equilibrium of the 

 atom becomes unstable, an explosion occurs resulting in 

 an expulsion of corpuscles, and such a shaking up of 

 those left in the atom that these vibrate so vigorously 

 that the energy radiated is sufficient to produce luminosity. 

 Thus I regard the ionisation of the gas as being due, not 

 to the corpuscles in the atom being dragged out by the 

 direct action of the electric forces in the field, or as being 

 knocked out by a rapidly moving corpuscle striking against 

 them, but to an explosion due to the atom having absorbed 

 so much internal energy that its equilibrium becomes 

 unstable. Other phenomena point to this as the method 

 by which ionisation is effected. If the corpuscles are 

 dragged out of the atoms by the electric field, the velocity 

 with which they are projected should depend upon the 

 strength of the field ; while if they are projected by an 

 explosion their velocity would depend only upon the nature 

 of the atom, and not upon the strength of the field. Now 

 when Rdntgen rays fall upon a substance the atoms of the 

 substance are ionised, and corpuscles are emitted forming 

 a stream of kathodic rays. Barkla has lately shown, how- 

 ever, that the penetrating power of the kathodic rays 

 produced in this way is independent of the intensity of 

 the Rontgen rays. Now the electric force in the Rontgen 

 rays depends upon their intensity, and the penetrating 

 power of the kathodic rays depends upon their velocity, 

 so that this result shows that the velocity of the corpuscles 

 does not depend upon the intensity of the force acting 

 upon them. Again, Lenard has shown that the velocity 

 of the corpuscles ejected when ultra-violet light falls upon 

 a metal is independent of the intensity of the light. 

 Lenard also investigated the secondary kathode rays pro- 

 duced when kathode rays fall upon matter, and found 

 that, in addition to rays the velocity of which was of the 

 same order as that of the primary rays, and which may 

 be regarded as deflected primary rays, there were other 

 very slow rays, and the measurements he gives indicate 

 that the velocity of these varies but little from that of the 

 primary rays. 



A point of great importance which can easily be shown 



NO. 1899, VOL. 73] 



by this apparatus is that the stage at which luminosity 

 sets in depends upon the current density through the tube, 

 and not merely upon the potential difference. One way 

 of showing this is to lower the temperature of the platinum, 

 keeping all the other conditions the same, and again 

 determine the relation between the current and the potential 

 difference. The effect of lowering the temperature is to 

 reduce the number of corpuscles starting from the kathode, 

 so that with the same potential difference the current 

 density is smaller. If the relation between the current 

 and potential difference is represented by a curve such as 

 Fig. 3, it will be seen at once that the lower curve cannot be 

 deduced from the upper curve by reducing all the ordinates 

 in the same proportion. The critical points on the curves, 

 i.e. the place where ionisation by collision begins and where 

 the luminous discharge appears, are at very different poten- 

 tials : the greater the current density the smaller the 

 potential difference corresponding to these critical points. 

 Thus, to take a case actually observed. When the wire 

 was very hot the discharge was brightly luminous with a 

 potential of 24 volts ; on lowering the temperature no 

 luminosity could be detected with a potential difference of 

 no volts. 



We can also show the effect of current density without 

 altering the temperature of the kathode by placing near 

 the tube an electromagnet so arranged that its lines of 

 magnetic force in the discharge tube are along the line 

 joining the kathode and the anode; the effect of the mag- 

 netic field is to make the corpuscles move along the lines 

 of force, and thus without altering the number of cor- 

 puscles emitted by the kathode it concentrates their paths 

 and so increases the maximum current density in the 

 tube. When the magnet is on, ionisation by collision and 

 luminosity both occur at a much lower potential difference 

 than when it is off, and it is easy to arrange matters so 

 that, keeping the potential difference constant, the discharge 

 is luminous when the magnet is on and dark when it is 

 off. When the potential difference is too small to produce 

 a bright discharge even when the magnet is on, the current 

 through the tube is often greater when the magnet is on 

 than when it is off. By placing the magnet so that the 

 lines of magnetic force are across the line joining the 

 kathode to the anode we can render the paths of the 



Discharge 



^rSSSMur Disc har^ 



'11 



U ° FJ£iO ZO 4CI e0 SoVolts 



Fig. 3. 



corpuscles more diffuse than they would be without the 

 field, so that the maximum current density is less when 

 the magnet is on than when it is off ; in this case it 

 requires a larger potential difference to produce a luminous 

 discharge with the magnet on than with it off. Similar 

 effects produced by a magnet on another kind of discharge 

 are described in my " Recent Researches," p. 105. 



