S'2 Mr. Hiebard>on on the Positive Ionization produced 



the E.M.F. were also made in hydrogen at a pressure of 

 •6 mm. In this case the apparatus was slightly different, the 

 spiral being parallel to, and about I'D cm. distant from, a flat 

 circular electrode to which the leak was measured. The 

 temperature was 692° C The relation between the current 

 in scale-divisions per half minute and the potential-difference 

 in volts is given in the following table : — 



Voltage. 



Current. • 



1 



40 



550 



80 



1300 ' 



120 



2600 



160 



3650 



200 



5150 i 



It will be seen that the relation between current and P.D. 

 here is not linear, but that the current increases more rapidly 

 than the voltaoe with high voltaoes. 



Further experiments were next made with the object of 

 testing the explanation of the discrepancy between the 

 ascendino- and descendino; curves shown in fior. 1. With this 

 aim the way in which the current under a constant voltage 

 varied with time was observed. The temperature was 

 maintained at 467° C. ; the pressure being "00625 mm. 

 Throughout the observations a potential of +40 volts relative 

 to the cylindrical electrode was maintained on the wire. As 

 in the previous experiments, there was no leak when the wire 

 was put to earth or when it was charged negatively. 



Corresponding values of the current in scale-di^-isions per 

 minute, and of the time in minutes from the commencement 

 of the experiments, have been plotted on squared paper (PL Y. 

 fig. 2). It will be seen that the observational points fall 

 very approximately on the curved line shown. The form of 

 the curve shows that the rate of decay of the current is great 

 at first, but after about an hour almost vanishes. The leak 

 then becomes constant, and approximately equal to 13 scale- 

 divisions per minute. For brevity we shall call the final 

 value of the current the *' steady leak '^ : the difference 

 between the value of the current at any time and the steady 

 leak may be called the '* induced '^ leak. The propriety of 

 usino- the term induced in this connexion will be demonstrated 

 later. 



We can explain these results if we suppose the induced 

 leak to be due to some substance which gradually decomposes 



