Atom and the Charge of Electricity carried by it. 541 



air in a vessel into which a point is discharging, we have to 

 consider rot only the communication of electricity from the 

 point to the gas, but also the diffusion of the electrified atoms 

 of the gas from the neighbourhood of the point through the 

 vessel. Any difference in the rate of diffusion would, even 

 though the communications of + and — electricity from 

 the point to the jar were made with equal ease, give rise 

 to an electrification of the gas. Imagine, for example, that 

 negatively electrified atoms diffuse through the gas more 

 quickly than positively electrified ones; then, if the rates of 

 emission of + and — electricity were the same, the negatively 

 electrified atoms would get further from the point than the 

 positive, so that the outlying regions would get a charge of 

 negative electricity ; the region round the point would be 

 charged with positive electricity, part of this would leak 

 back into the point, leaving the gas as a whole negatively 

 charged. Several phenomena, which I shall now proceed to 

 describe, have led me to the conclusion that the effects due 

 to the communication of the electricity from the metal to the 

 gas are masked in many cases by the difference in the rates 

 at which the electrified atoms diffuse, the negative atoms in 

 all my experiments diffusing more rapidly than the positive 

 ones. In the case of oxygen, where the gas gets a positive 

 charge, the strong tendency for the oxygen to combine with 

 the metal is sufficient to overcome the more rapid diffusion, 

 whereas with such gases as hydrogen and carbonic acid, where 

 the tendency to combine is weaker, diffusion gets the upper 

 hand. 



I made several experiments with chlorine, using dif- 

 ferent materials for the discharging point. The first sub- 

 stance I tried was silver chloride which had been fused into 

 a solid mass; when this point was connected to one of the 

 terminals of the high tension transformer, the chlorine was 

 negatively electrified. I then used a zinc point, protecting 

 the zinc before the experiment began by covering it with a 

 glass tube, through which it was pushed just before the dis- 

 charge began ; in this case, again, the chlorine was negatively 

 electrified to a slight extent. I then replaced the zinc point 

 by a sodium one, protecting this in the same way as the zinc 

 by a glass tube ; when the discharge passed from the sodium 

 point, the chlorine was positively electrified. Thus by using 

 a substance such as sodium, which has a very great affinity 

 for chlorine, we are able to overcome the effect of diffusion, 

 and get a positive charge in the gas, just as when the dis- 

 charge passes from a copper point into oxygen. 



The following experiments seem to show that a negatively 

 Phil. Mag. S. 5. Vol. 40. No. 247. Dec. 1895. 2 P 



