404 Mr. K. H. Kino-don 



O ' 



on some 



Hence the probability that for any collision the relative 

 velocit} r of the molecules normal to the sphere of action shall 



be less than v is (l—e~i hmv2 ), since nV\/^- is the total 



V km 



number of collisions per c.c. per second. Similarly the pro- 

 bability that the relative velocity tangential to the sphere of 

 action shall be greater than u is e -i hmu2 m Therefore the 

 probability that both these conditions are fulfilled for a par- 

 ticular molecule is 



and the number of such collisions per c.c. per second will be 



N" = n 2 a 2 1 / ^ZL e -* hmu - ( 1 — e -V™*>\ _ 

 V km 



Experiments. — The only experimental work which has 

 been done on this subject is by Patterson *, and by Devik f. 

 In Patterson's work the gas was contained in an iron cylinder, 

 and as iron usually contains some radioactive impurity, the 

 number of ions generated per c.c. per second was quite large 

 (n = 61). He failed to detect any effect of temperature on 

 the ionization up to 400° C, but it is possible that the effect 

 might have been masked by the largeness of the currents 

 measured. Also, as the air in the receiver was always at 

 atmospheric pressure, its density would decrease as the tem- 

 perature was raised ; this decrease in density would decrease 

 the ionization current due to the earth's penetrating radiation 

 and also that due to a radiation of the /3 or <y type coming 

 from impurities in the walls of the chamber, both of which 

 form part of the total current measured. 



In Devik's experiments the gas was momentarily heated 

 by an adiabatic compression, and the ionization measured at 

 the moment of greatest compression. The only gas which 

 showed any signs of ionization caused by the high temperature 

 (estimated at 900° C.) was antimony hydride. 



In view of the methods of the above experiments, it was 

 thought worth while to carry out another investigation in 

 which the following conditions should be satisfied : — (1) the 

 ionization-chamber should be airtight : (2) the residual ioni- 

 zation should be as low as possible so that any change would 

 make itself more apparent; (3) the temperature should be 

 kept constant during the time of each reading. Unfortu- 

 nately, in order to fulfil these requirements the range of 



* Patterson, Phil. Mag. vi. p. 231 (1903). 



t Devik, Sitz. d. Held. Ahad. Whs. xxiv. (1914). 



