the Electrical Conduction produced by it. 



155 



The ends of the cylinder D were closed by ebonite collars, 

 and the central cylinder was supported in position by brass 

 rods passing through the ebonite. The surface was uniformly 

 covered with uranium oxide. The cylinder D was connected 

 to one pole of a battery, the other pole of which was to earth. 

 The cylinder C was connected to the electrometer. The 

 following tables show the results obtained for Irydrogen, carbonic 

 acid, and air. Distance between cylinders *375 cm. 



Hydrogen. 



Carbonic Acid. 



Air. 



Volts. 



Current. 







122 



-•062 



125 



•125 



123 



•25 



142 



•5 



150 



1 



160 



2 



163 



4 



165 



8 



1C8 



16 



172 



108 



178 



216 



185 



Volts. 



Current. 







95 



-•125 



205 



•25 



255 



•5 



305 



1 



355 



2 



405 



4 



460 



8 



520 



16 



590 



36 



705 



108 



787 



216 



820 



Volts. 



Current. 



I 



+ 1 



418 



2 



451 



4 



495 



8 



533 



36 



601 



108 



615 



216 



630 



The above results are expressed graphically in fig. 13, 

 where the ordinates represent current on an arbitrary scale 

 and the abscissse volts. In the tables given for hydrogen and 

 carbonic acid it will be observed that the current has a 

 definite value when there is no external electromotive force 

 acting. The reason for this is probably due to the contact 

 difference of potential between the uranium surface and the 

 interior brass surface of the outside cylinder. When the 

 external cylinder was connected to earth the inside cylinder 

 became charged* to —'12 volt after it was left a short time. 



* This phenomenon has been studied by Lord Kelvin, Beattie, and 

 S. de Smolan, and it has been shown that metals are charged up to small 

 potentials under the influence of uranium radiation. The steady 

 difference of potential between two metal plates between which the 

 radiation falls is the same as the contact difference of potential. An 

 exactly similar phenomenon has been studied by Perrin (Comptes Rendus, 

 cxxiii. p. 496) for #-rays. 



