of Secondary Cathode Rays produced by X-Rays. 43 



Since \i is small compared with /?, the most important part 



of this equation is Re 2 . Designating the other factors by 

 S we may write 



R= e/i 



be - 



If we assume that the absorption coefficient of the cathode 

 rays is proportional to the density of the absorbing material, 

 we can calculate the values of /3 from the values given by 

 Sadler * for their coefficient in air. 



Sadler f has measured the coefficients of absorption in air 

 of the fluorescent radiations characteristic of copper and 

 arsenic, and he and Barkla J have shown that the ratio of the 

 coefficients of any two fluorescent beams is constant for all 

 absorbing substances in which these beams do not excite a 

 fluorescent radiation. We can, therefore, calculate the values 

 of X 2 in air from those in aluminium and in air given by 

 Barkla and Sadler. 



The fluorescent rays used were those from chromium, iron, 

 zinc, and tin, and with the exception of tin Barkla and 

 Sadler § have measured their absorption coefficients in gold 

 and silver. Their value for the coefficient of the rays from 

 chromium in gold is unreliable according to their statement, 

 but it will be seen that \i is always so small with respect to 

 (3 that an exact knowledge of this quantity is not important. 



The absorption coefficients of the rays from tin in gold and 

 silver are not given by the above writers, but may be 

 obtained approximately from the constant ratio between 

 absorption coefficients cited above. Tin rays, being more 

 penetrating than those characteristic of silver and gold, will 

 excite their fluorescent radiations and, therefore, will probably 

 be less penetrating to these metals than the above calculation 

 would indicate, but an inspection of the curves given by 

 Barkla will show that the increase of absorption is not very 

 great in this part of the spectrum, and it will be seen that 

 the effect of \ l is negligible in the case of the rays from tin. 



The accompanying tables give the values of the quantities 

 occurring in the formula as calculated from the data given 

 by Barkla and Sadler and the observed values of e/i in gold 

 and silver. The last column gives the values of R as calcu- 

 lated from these results. Most of the values of e/i were 

 obtained using only one sheet of metal foil, but in some cases 



* Phil. Mag. vol. xxii. p. 447 (1911). 

 t Loc. cit. 



X Phil. Mag. vol. xvii. p. 739 (1909). 

 $ Loc. at. 



