t08 //. A. Bumstead — pj>ii».sion of Electrons by 



omit ted by the case under the influence of the a-rays as from 

 the two sides of the aluminium foil electrode through which 

 the rays passed. This difference persisted when the case itself 

 was lined with aluminium foil, so it was not due to a specific 

 difference between aluminium and brass. It is probable that 

 the explanation of this difference, as well as the difficulty in 

 obtaining negative saturation mentioned in the last paragraph, 

 is to be found in the construction of the cover and the aper- 

 tures through which the rays enter the exhausted chamber. 

 Many of the rays, passing in a divergent pencil through the 

 aluminium foil which covers the holes, strike the sides of the 

 holes and do not reach the electrode. They thus liberate a 

 number of electrons from the case and no corresponding ones 

 from the electrode ; and to draw all of these superfluous elec- 

 trons from the small apertures in the top of the cover requires 

 a considerably greater potential-difference than when they come 

 from the electrode. 



In order to find out whether the observed currents really had 

 their source in the metals and were not due to residual gas or vapor, 

 the foil was removed from the ring electrode. (The apertures 

 in the top of the cover were so placed that the geometrical beam 

 of rays fell entirely within the ring.) Under these conditions 

 the current with the case charged positively fell to 4 per cent 

 of its former value, while with a negative potential on the case 

 the current was 90 per cent of what it had been before. This 

 slight falling off in the negative current was doubtless due to 

 the fact that the ring-electrode was less efficient in catching 

 electrons from the case than when it was covered with foil. 

 Of the small positive current, part at least was due to the rays 

 which struck the brass rod which supported the ring (see fig. 1), 

 so that we may be quite sure that the gas effect contributes not 

 more than 2 or 3 per cent to the currents observed with this 

 apparatus. The absence of positive particles justifies the fur- 

 ther conclusion that the effect produced by the rays on a metal 

 consists in the direct expulsion of electrons, and not of neutral 

 pairs which afterward break up. As will be seen later, there 

 is a very close relationship between this metal effect and the 

 ordinary ionization of gases by a rays ; and the result just given 

 renders it at least unlikely that neutral pairs are emitted when 

 gases are ionized by a-rays as has been sometimes assumed in 

 the case of other ionizing agents. 



A further test of the relative importance of the metal and 

 gas effects was made by measuring the current at different gas 

 pressures up to 0'2 mm . The result may be expressed by the 

 linear formula 



i = A + Bp 



