794 



SCIENCE. 



[N. S. Vol. V. No. 125. 



at diflferent densities of the gas. Thomson 

 and Eutherford {Phil. Jfa^., 42, 392) and 

 others have already studied this subject 

 and shown that the current does not obey 

 Ohm's law. As high E. M. F.'s are reached 

 the current does not increase in a manner 

 proportional to the increase in E. M. F. In 

 Fig. 2 I have plotted such a set of curves, 

 and it will be noticed that the form of the 

 curve is not the same for different densities 

 of the gas. Thomson and Rutherford have 

 shown that the form of curve is not the 

 same in the cases of air and hydrogen. The 

 curve for air as given by them (Ibid, p. 

 404) is similar to the curve given in Fig. 2 

 for normal density of the air. That for 

 hydrogen is similar to that given in Fig. 2 

 for pressure of 250 mm. However, it can 

 scarcely be said that the form of the curve de- 

 pends upon the density of the gas. The curve 

 for mercury vapor as given by Thomson 

 and Rutherford does, indeed, differ from that 

 given for air in the opposite manner from 

 that for hydrogen, but their curves for 

 chlorine, sulphuretted hydrogen and air 

 coincide in form, and the densities of these 

 gases diifer widely. They infer from their 

 theory of this action " that more conduct- 

 ing particles are produced by the rays in 

 air than in hydrogen, but that the product 

 of U, the velocity of these particles, and T, 

 a time which is proportional to the time 

 these particles linger after the rays are cut 

 off, is greater for hydrogen than it is for 

 air." It is altogether possible that the same 

 explanation may apply to the case where 

 the same gas is used at different pressures. 

 However, experiments to prove this would 

 certainly be desirable. They also state that 

 "the gases which have large saturation 

 curves are those which contain the elements 

 which have abnormally large specific in- 

 ductive capacities in comparison with their 

 valency." This remark does not seem to 

 apply to the case where the same gas is 

 used at different pressures. The saturation 



currents here do differ, and it is hard to 

 see how either the specific inductive capaci- 

 ties or the valencies can differ for air at dif- 

 ferent densities. 



It has also been found that the form of 

 the curve between rates of discharge and 

 density of air is different for different, 

 intensities of the rays. The curves which 

 have already been plotted were taken with 

 the tube near to the plate to be discharged. 

 A series of curves were also taken with the- 

 tube at some distance from the plate and 

 several inches of board placed between the 

 two . These curves were somewhat similar to 

 those shown in Fig. 1, but there was not as 

 marked a difference between the curves for 

 different potentials. In fact there was no 

 curve which showed a maximum for pres- 

 sures less than atmospheric pressure. 

 It, therefore, seems scarcely advisable 

 to try to get the form of the curves with 

 any great degree of accuracy at present, for 

 the form depends upon the intensity of the 

 rays and there is no way of determining 

 this intensity in any standard unit. I can 

 define the intensity no better than I have 

 already defined it in stating the rate of dis- 

 charge which it produces from a zinc plate 

 of given area with an approximate guess at 

 the number of units of capacity in the sys- 

 tem of plate, condenser and connections. 



It has been stated by Perrin (C. R. 124, 

 454) that the discharge effect can be sepa- 

 rated into a surface effect between the gas- 

 and the charged metal, and a volume effect 

 throughout the gas. It was, therefore, 

 thought that covering the zinc plate with a 

 thin film of parafBn might change the form 

 of the curve when the density of the air is 

 changed. This was tried and no such change 

 was noticed. It has not been possible to 

 keep the action of the tube exactly constant, 

 and so one cannot be entirely sure of the 

 correctness of the results, but if coating the^ 

 zinc plate with parafBn causes any change- 

 in the form of the curve it is at least small. 



