MK. j. s. TOWNSEND ON THE DIFFUSION OF IONS INTO GASES. 143 



When it was required to find the conductivity of air after passing through nhr.rt 

 tuljes, the tubes T, were removed and the disc y was placed in the position occupied 

 by at, then the electrode was moved up near the disc so that the electric force should 

 act on the air immediately after leaving the tubes, T 2 . 



Experiments with oxygen, hydrogen, and carbonic acid were made with the 

 apparatus, the horizontal section of which is shown in fig. 2. It consisted of two 

 long tubes, A, and A 8 , each exactly similar to A in fig. 1. In one of them the long 

 tubes, T|, were set up, and in the other the tubes T 2 . The tubes, G and H, were 

 connected to two gasometers, so that the gas could be passed from one to the other, 

 through either of the tubes, A, or A,. The two tubes were fixed tightly into two 

 rectangular pieces of ebonite, R and It', which rested on the top of the box containing 

 the bulb. Two wooden rails we're screwed to the box at such a distance that the 

 ebonite supports fitted exactly between them, so that, by sliding the apparatus from 

 side to side, the window in either of the tubes could be brought exactly over the bulb. 



It was found necessary to put a cylinder of aluminium inside each of the tubes, 

 extending from p to p', to prevent the rays from falling on the inner surfaces of the 

 tubes, which were of brass. Before these cylinders were put in experiments were 

 made to see whether the ionization produced in a stream of air passing along A, was 

 equal to that produced in an equal stream through A 2 , and it was found that there 

 was a considerable difference between the conductivities in the two cases. The 

 inequality was not due to any difference in the thickness of the aluminium covering 

 the two windows, but was traced to differences in the state of the surfaces of the 

 brass tubes opposite the windows. 



It has been shown by PERRIN* that the ionization produced by Rontgen rays in a 

 gas in contact with a metal is considerably increased by allowing the rays to fall 

 normally on the metal surface. This effect upon the ionization is different for 

 different metals, and depends also upon the state of the surface. According to 

 PERRIN, only a very small increase in conductivity is produced when the rays fall 

 upon an aluminium surface. It was found that, when the two aluminium cylinders 

 were put inside A, and A 2 , the difference in conductivity which was first observed 

 disappeared entirely. 



METHOD OF CONDUCTING THE EXPERIMENTS. 



When working with the first form of apparatus the experiments were conducted 

 in the following manner : The tube A is raised to a potential of 80 volts positive, 

 and the quadrants to which the electrode is joined are insulated. The stream of air 

 from the gasometer is thus allowed to pass through the apparatus, and, when the 

 velocity is steady, the coil working the bulb is turned on for a fixed time (20 seconds 

 generally) and a deflection of n, divisions is obtained on the electrometer scale. 



* ' Comptes Rendus,' vol. 124, p. 455. 



