and the Constitution of the Atom. 2^1 



For small atomic numbers we can put e = and 

 ^/R = 3/4N 2 — BN + C Calculating j//R for various values 

 of q 2 we should find that q 2 = 9 gives the best agreement 

 with observations for small atomic numbers ; but also the 

 other values of q 2 give a fairly good agreement, and, as 

 already mentioned in the general discussion of the problem, 

 q 2 must be determined from the L-radiation. 



In Paper I. we found that q 2 =7 and q 3 =l would give a 

 good agreement in the case of L a . As we shall see later on, 

 the explanation of L a on the assumption of recombination 

 between primaries and conservation of momentum gives 

 q 2 = l. 



Thus we put : 



qi = o, q 2 = 7. 



In accordance with Debye we introduce for the sake of 

 convenience the quantity A given by the expression 



A=g-3/4N 2 =-BN + C + e. 



In the present case we have 



€ = 1-325 10- 5 J3(N-0-577) 4 -2(N--0-25) 4 - ^(N-4'305) 4 



+ 



r(N-4-827)4 



In Table IV. are given values of A corresponding to 

 various values of N. 



Table IV. 



N... 11 15 J20 25 30 35 40 45 50 55 60 

 A ...-17-2 -24-4 -32-6 -39*5 -44-4 -462 -436 -35-5 -20-4 ■+ 51 +387 



In fig. 1 is drawn the curve showing the variation of A 

 with the atomic number. The observed points are also 

 marked off. For the sake of comparison two other curves 

 are drawn, that of Debye (<?i = 3, <? 2 = 1) and the one corre- 

 sponding to maintenance of energy (^ 1 = 4, ^, = 3). (See 

 Paper I., p. 308.) 



We see that the curve of Debye is the best for small 

 atomic numbers ; but the two other curves give a better agree- 

 ment for high atomic numbers. From a mere numerical 



