246 Mr M c Clelland, On the Conductivity of Gases 



as abscissa?. The curve cannot be produced back to the arc, as 

 that part includes the glass funnel, where the velocity of the 

 stream of gas was different. 



The loss of conductivity is principally due to the recombination 

 of the carriers, and if we assume the number of collisions of the 

 positive and negative carriers to be proportional to the square of 

 the number present, we get 



dn 



= an 



dt ' 



Nn 1 



01 W^Ti ~ at ' 



where iV is the number of carriers when t — 0. 



If we take N= 86 and measure t from that point and calculate 

 1/a for each of the other readings we get 



1/a proportional to 306 



» » >> 301 



„ 311. 



The curve therefore agrees fairly well with that got by assuming 



dn 

 — 77 = an 2 . 

 dt 



Determination of the velocity under an electric force of the 

 carriers produced by the arc in air. 



(5) The apparatus shown in Fig. 1 enables us to measure the 

 rate at which the carriers of electricity move in an electric field 

 of given strength. 



The rate of leak from B is first measured when G is to earth, 

 and B charged sufficiently high to discharge all the carriers before 

 they pass it. The rate of leak from B is again determined when 

 G is kept at such a potential that the leak from B is diminished 

 by about one-half. Knowing these two rates of leak, the potential 

 of G, and the velocity of the stream of air past the terminals, we 

 can calculate the velocity of the carrier. 



If V be the potential of G, v the velocity of the carrier under 

 a force of 1 volt per cm., r x and r the radii of the tube A and the 

 terminal G respectively, then we can easily show that all the 

 carriers of one sign initially inside a radius p are discharged to C 

 where p is given by 



1 p 2 — r,, 2 . r x 



