Velocity and Rate of Recombination of the Tons of Gases. 423 



apart along the brass tube. The brass tube was connected to 

 earth, and one of the electrodes was connected to the electro- 

 meter. Both quadrants were connected together, and the 

 whole was charged up to a high potential (in practice about 

 100 volts). The quadrants were then insulated from each 

 other. When the rays were turned on, there was no effect 

 on the electrometer until a current of air from the gasometer 

 passed along the tube, when the movement of the electrometer 

 needle showed that the air passing the electrodes was con- 

 ducting. The rate of leak for each of the electrodes was 

 tested in turn, and since the capacity in the electrometer 

 circuit was nearly equal in the three cases, the rates of leak 

 were proportional to the conductivity of the air at the 

 electrodes. 



By noting the volume of air which passed from the gaso- 

 meter in a given time, and knowing the diameter of the 

 metal tube, the mean velocity of the current of air could be 

 readily calculated, and therefore also the intervals of time 

 taken by the current of air to pass from" one electrode to 

 the other. By altering the weights on the gasometer the 

 velocity of the blast of air could be varied at will. 



The following table shows- the way in which the conducti- 

 vity of the air varied with the length of time after exposure 

 to the rays. The first column gives the times taken by the 

 current of air to pass from one electrode to the other, and the 

 second column gives the ratio of the rates of leak of the 

 electrodes. 



Time Ratio of the 



in seconds. rates of leak. 



•13 -75 



•22 -61 



•28 -57 



•65 -39 



24 -11 



It will be seen that the conductivity falls off rapidly with 

 the time, and after 2*4 seconds is only ^ of the original 

 value. 



It is probable that some of the conducting particles give 

 up their charge to the sides on their passage down the tube, 

 but the correction for this is probably very small as there is 

 no force acting on the charged gas tending to repel itself to 

 the sides as is the case when free electrification is present. 



In a previous paper (J. J. Thomson and E. Rutherford, 

 Phil. Mag. Nov. 1896) it has been shown that when a gas is 

 acted on by the Rontgen rays a steady state is reached when 



