Absorption Coefficients of X Rays. 629 



fact is that the rays can excite the characteristic rays o£ the 

 former group but not of the latter, and spend themselves in 

 doing so, and in exciting at the same time large numbers of 

 cathode rays. The shortest wave of the six is just short 

 enough to excite secondary rays in Pd, and is consequently 

 highly absorbed by that substance. 



Although the figures in each column are not all the same, 

 the want of order in their irregularities shows that experi- 

 mental errors are the cause of the lack of uniformity. 

 Barkla's rule, that the ratio of two absorption coefficients 

 is independent of wave-length over considerable ranges, is 

 fully maintained. 



That being so, we may average the differences in each 

 column, and expect thereby to obtain greater accuracy. 

 The averages are put in the last row but one. 



We may now look for some relation between the atomic 

 absorption coefficients and other atomic characteristics. It 

 is natural to seek first for some connexion with the atomic 

 numbers which have lately been shown by Moseley to be so 

 fundamental. It appears at once that over considerable 

 ranges the coefficient is proportional to the fourth power of 

 the atomic number. In order to make this clear the last 

 row of Table III. contains a set of figures each of which is 

 four times the difference of the logarithms of the atomic 

 numbers of the substances belonging to that column. For 

 instance, the atomic numbers of Fe and Al are 26 and 13; 

 and 4 log (26/13) is equal to 1 201. 



The close parallelism of the figures of the last two rows 

 implies that the absorption coefficient of any substance of 

 atomic number N is equal to CN*, where C has constant 

 values over prescribed ranges. For the wave-lengths under 

 consideration the constant changes its value suddenly in 

 passing through the Rh-Pd-Ag group. The constant is a 

 function of the wave-length. Owen * has shown that the 

 absorption coefficient is inversely proportional to the fifth 

 power of the atom of which the X ray is characteristic, the 

 absorbing substance remaining the same. Darwin has re- 

 stated the rule in accordance with the newer modes of 

 expression, and puts it in the form 



Absorption coefficient oc (\vave-length) 5/2 . 



The results given above are in fair accordance with this 

 rule, but their range is too small to make a satisfactory test. 



* Proc. Roy. Soc. 1912. 



