the Absorption of X-Ray$. 189 



not possible for elements which show greater absorption. 

 Already at Si this method must be regarded as less reliable. 

 The values for Si just mentioned plausibly authorize the 

 assumption that the number of ihe outer electrons must be 

 less than 6. There is not a long way to the suggestion 

 that the atomic structure of Si is analogous to that of 0, 

 which, indeed, in the periodic system stands next above Si in 

 the same vertical row. Consequently, if we suppose 4 outer 

 electrons, the values given in the table will be obtained, 

 which may be regarded as almost constant. In the Si-atom, 

 just as was the case in C, 2 electrons would be more closely 

 attached to the nucleus than the 4 outer electrons. Thus we 

 should be able to account for the value of ^ Si/Cll being some- 

 what higher than it ought to be if this element, like the next 

 preceding and following elements, only contained an inner 

 region with 8 electrons. From the sudden increase of 

 absorption at S (Z = 16) apparently a new rearrangement 

 takes place at this element, and the above reasons speak in 

 favour of the assumption that here 4 electrons pass from the 

 inner region, which thus should contain 12 elections. The 

 capacity of absorption indicating that the arrangement of 

 electrons in the inner region of the elements 01 (Z = 17) 

 and K (Z = 19) remains unchanged, these atoms would 

 contain 5 and 7 electrons respectively in the outer region. 

 At Ca (Z = 20) there would occur another rearrangement of 

 the electrons, by which change four other electrons would 

 pass from the outer to the inner region. Thus the inner 

 region of the elements Ca-V would contain 16 electrons 

 and the outer region, as in the preceding group, 4-7 electrons 

 respectively. p 



With increasing atomic weight the term - — ^- b becomes, 



Jc Cu 



in comparison with the term ~- (5), of still less importance 



Cu 

 and, certainly in the latter of the just-mentioned elements, 



lies within the limits of experimental errors. If, however, 



the relation between 7e a /c u and Z is pictured graphically 



(fig. 2), it is seen that /e a /c u within each group appears to be 



very nearly a linear function of the atomic number. Thus 



the relation between the two quantities may be expressed : 



* a/ O»=K(Z-<0, (7) 



where c stands for the number of inner electrons of the 

 elements constituting the group, and K is a constant cha- 

 racteristic of the group. If the straight line designating 



