Ions produced by Rontgen Rays. 11 



For short intervals of time, however, there is a marked 

 curvature indicating that a is relatively large for the initial 

 stages of recombination, but diminishes, becoming in most 

 cases nearly constant for t greater than about one-third of a 

 second. The curve for air is nearly straight at the end. In 

 the case of carbon dioxide the curve is surprisingly steep at 

 the end, indicating a very large value of a as compared with 

 air, and one would suppose that it must ultimately become 

 more nearly horizontal. It was therefore desirable to further 

 extend the curve for air and carbon dioxide. In order to do 

 this the pendulum was arranged to operate the keys on the 

 return stroke. A key operated by hand was put in parallel 

 with the key which applied the field and held down by hand 

 until the pendulum was approaching the latter on the return 

 stroke. The pendulum key was then quickly closed and the 

 hand key released. 



In this way it was possible to extend the air and carbon- 

 dioxide curves to t = l'06 seconds. The air curve was found 

 to bend very little more, bat the carbon-dioxide curve became 

 finally only a little steeper than the air curve. 



Discussion of Results. 



The high initial values of a and the subsequent falling off 

 of its value, first rapidly and then more slowly, may be 

 explained as follows : — 



If we suppose the action of Rontgen rays is the liberation 

 of rapidly moving electrons from the gas molecules, which 

 in turn liberate slow moving electrons from molecules in their 

 path, thus ionizing them, one would expect the ions to be 

 segregated in the path of the rapidly moving electrons as 

 shown by C. T. R. Wilson's cloud experiment. This would 

 make the value of a greater than for a uniform distribution 

 of ionization and as a result of diffusion the value of a would 

 slowly diminish. This would probably not account for the 

 very high initial values of a. It has been found * that the 

 /3 particle produces on the average only about 67 ions per 

 centimetre of path. As a rule, therefore, there must be a 

 very large number of molecules between them, and the 

 distance between the two ions constituting an initial pair 

 would be in general less than the average distance between 

 the ions strewn along the path of the ft particle. 



If a similar condition exists in the case of ionization by the 

 electrons expelled under the action of X-rays, there would be 

 a relatively great probability of collision between the positive 



* Geiger and Kovarik, Phil. Mag. (6) xxii. p. 604 (1911). 



