1920. No. 2. ON THE X-RAY SPECTRA. 45 



Cg. The curve of Bf.nkdicks (Paper I. fig. 3) shows a gradual increase 

 of the conductive capacity indicating an increase of the quantnumber. 

 From the A'-ray analysis we know that when we come to elements, 

 which give the L-radiation there is a second ring with quantnumber 2, 

 which has 7 electrons. 



This result might indicate that the surface ring of the elements of 

 the first short period had the quantnumber 2 — , but we must also be 

 aware of the possibility that some change of the quantnumber of the 

 second ring may have taken place, or we cannot be sure that also the 

 elements before Krypton have a second ring of 7 electrons with quant- 

 number 2. The very large increase of electric conductive capacity from 

 Li to Na would be best accounted for by assuming that the surface 

 electron of Li had quantnumber i. In that case the conductive capacity 

 of the elements considered should vary as the numbers: i. 4, 9, 16, 25. 



These numbers give at any rate variations of the right order of 

 magnitude. In fact the conductive capacity of C^ is about exactlv 

 25 times as large as that of Li. 



But we ought not lay too much stress on this coinsidence, for the 

 conductivity may be influenced from many other causes. 



Putting the quantnumber for the surface electron of Zi equal to 2, 

 the conductive capacities should vary as the numbers 4, 9, 16, 25, 36. 

 It should be remembered that the constitution given to Fe and Ni in 

 deducing the values of v^ for the Ä'-absorption edge, was in accordance 

 with the latter assumption. 



A special interest is attached to the elements Ca, Ag, Au, which 

 have a large conductive capacity and only one electron in the surface 

 ring. We notice a very marked increase from Cu to Afj corresponding 

 to an increace of quantnumber; but for Au there is again a drop of 

 conductive capacity, which might indicate a drop of quantnumber. It 

 seems as if the formation of the rare earth elements is accompanied with 

 some kind of condensation process. 



The atomic property which first of all suggests itself for examination 

 is the atomic volume, because we might expect it to vary in a similar 

 way as the radius of the surface ring in such a way that a greater radius 

 would give a greater atomic volume. 



Now the radius of the surface ring is at any rate approximately 

 determined by the equation (30) and by means of the values previously 

 given to q and n of the surface ring we can easily calculate the radius 

 ft and compare its variation with the atomic volume curve. 



