430 Mr. E. Rutherford on tlie Velocity and Rate 



the value of the after-conductivity was taken next day without 

 disturbing the gas. 



The quantity of electricity that passed through the gas 

 after the rays had ceased gave a deflexion in the electro- 

 meter of 70 divisions, and the value of T was 1 second. A 

 blast of dusty air from a bellows was then sent into the 

 cylinder and the deflexion due to the after-effect fell imme- 

 diately to 15 divisions, with a value of T of about *15 sec. 

 When the air was allowed to stand the after-effect gradually 

 increased again to 35 divisions, with a value of T of '5 sec. 

 after an interval of about 10 minutes. Several hours elapsed 

 before the after-effect rose to 60 divisions. This experiment 

 shows what a variable quantity T is for the same gas, de- 

 pending as it does on the amount of suspended matter in the 

 gas. 



The effects observed in air and other gases seem to point 

 to the conclusion that freely suspended particles greatly assist 

 a gas to lose its conducting property after the rays have 

 ceased. 



Since the dust-particles are very large compared with the 

 ions, an ion is more likely to strike against a dust-particle, 

 and give up its charge to it or to adhere to the surface, than 

 to collide with an ion of opposite sign. A positive ion 

 striking a dust-particle gives it a positive charge, and this is 

 neutralized by a charge from a negative ion, and in this way 

 the rate of loss of conductivity is much more rapid than if 

 the loss of conductivity were due to collisions between the 

 ions themselves. It seems probable that if a gas could be 

 obtained completely dust-free the rate of recombination which 

 would be due entirely to molecular collisions would be very 

 much slower than for ordinary air. 



When the rays act upon a gas the number of ions per c.c. 

 increases until a definite stage is reached, when the rate of 

 production is equal to the rate of recombination. It is of 

 interest to find the time that elapses after the radiation has 

 commenced before this maximum is reached. In most of the 

 experiments there were generally 50 breaks per second in the 

 induction-coil, so that for the sake of calculation we may very 

 approximately suppose that the bulb was giving out rays 

 uniformly, corresponding to the production of q ions per c.c. 

 per second. The rate of increase of n is given by 



dn , 



-rr-=:q — an . 

 dt * 



Solving this equation it is easily seen that the time t required 

 for the production of n ions per c.c. in the gas is given by 



