﻿392 Coefficient of Absorption of the /3 Rays from Uranium. 

 radiation from a thick plate o£ the substance by the relation 



. __ l-P- A /l-& 



2' 



k = 



or if we denote R/I by the letter p 



"We are thus enabled to calculate fju when p and X are known. 



The secondary radiation in the case of uranium rays was, 

 however, so small, that the errors of observation were so great 

 as to outweigh any theoretical advantage there might possibly 

 be in the use of the " true " coefficient p, instead of the co- 

 efficient X. The return radiation R, was obtained as the 

 difference of two quantities, I and I + R, which in most 

 cases were nearly equal. The proportionate error in R 

 therefore was very great. Thus even in the case of lead, an 

 error of only 1 per cent, in measuring the intensity of I , or 

 I 4-R, would produce an error of 7 per cent, in the value of 

 p, while in the case of copper the error would be as much as 

 25 per cent. As, owing to the smallness of the radiation 

 from uranium, it was certainly not possible to ensure an 

 accuracy greater than 1 per cent, in the measurement of 

 the intensities of the rays, it was considered preferable to 

 retain the directly measured coefficient X. 



No great amount of error is introduced in this way, and 

 the general results of the experiment are quite unaltered. 

 In the majority of cases the secondary radiation was almost 

 inappreciable, and X therefore very nearly equal to jju. Sub- 

 stituting the values of p actually obtained, we have for alu- 

 minium, p, is equal to X ; for copper, p, equals 1*08 X ; while 

 for tin, gold, platinum, and lead, which show the i maximum 

 effect, fi is equal to 1*3 X. The effect, moreover, is roughly 

 proportional to the atomic weight. 



The effect of substituting fi for X in the curve of fig. 3 

 would therefore not affect the general appearance of the 

 curve, nor the general results deduced therefrom. Its only effect 

 would be to slightly increase the slope of the later portions, 

 and thus somewhat increase the rate of increase of the 

 absorption with atomic weight. 



In conclusion I wish to express my best thanks to Prof. J. J. 

 Thomson for his kindness in suggesting the subject of this 

 research, and also for much kindly interest during the course 

 of the experiments. 



Cavendish Laboratory, Cambridge, 

 August 3rd,, 1906. 



