44 L- VEGARD. M,-N. Kl. 



where 71 is the quantnumber and 7 the number of electrons of the sur- 

 face ring. 



Now we found that in order to get the observed typical variation 

 of the conductive capacity with atomic number (Benedicks curve) we had 

 to assume the quantnumber )) to be constant for all elements belonging 

 to the same period, or for all elements falling between two sucessive 

 inert gases. On this assumption we got the right type of variation of 

 the conductive capacity for both the short periods. (See Paper I. fig. 4). 

 But also for the two long periods {Ar — Kr) a.nd {Kr — Xc) we get the 

 right type of variation, when we assume n constant inside each interval. 

 Let us e. g. consider the period between Ar and Kr. At first we get a 

 sudden fall of ^/a with increasing j) as we pass from K to Ki. Then a 

 new ring is formed which should produce a sudden increase of ^ja 

 for the element Cu. In fact we find a sudden increase of the conductive 

 capacity for Cii followed by a fall which is continued till the next period 

 sets in with Kr. 



There is, however, an apparent disagreement with regard to the 

 magnitude of the conductive capacity. 



The formula (29) would give both for K ane Ca the same value of W, 

 while the observed conductive capacity is very much smaller for Cu. I 

 think, however, that the smaller value in the case of Cu is just what we 

 may expect from theory. The expression of i/rx given in equation (29) 

 is deduced on the assumption that the effect of the internal electrons 

 is the same as if they were placed at the centres. 



Now the radius of the surface ring is determined by the equation 

 (Paper I equation 25): 



« = «hY^ ■•■(30) 



where «^^ is the radius of the surface ring of Hydrogen. When we pass 

 an inert gas — Ar say — we shall have an increase a from the two causes: 

 increase of )) and diminution of q. When however, we pass from Xi to 

 Cu we have relatively smaller increase of a because now the quant- 

 number is unaltered. Now this comparatively much smaller jump with 

 regard to the radius which is found for Cii may account for its smaller 

 conductive capacity as compared with that of K. 



For chemically'' related elements q is the same, and the electric 

 conductive capacity as mentioned in Paper I., should vary in a similar 

 way as «-. 



Let us e. g. consider the elements for which q= i, and w^hich follow 

 immediately after an inert gas. These elements are Li, å^j, K, Rf^ and 



