306 Prof. W. M. Hicks on the 



m = o. The only directly observed line is D n . The 

 deduced lines 



(8) (1)28519*19 =28504*39 (1-6), (1) 30846-39 .£ = 34617-39 (-5) 

 can only be taken as numerical relations. 



??i = 6. The D lines are not observed but the values 

 calculated from the formula are inserted. With these we 

 get a directly observed D 22 , with a D n deduced as the mean 

 from the two linked lines 



e.{2) 36652-99 = 32881-99 ; (1) 30422*79 .«= 32881-43. 



As both the latter practically agree the allocation is very 

 probable and again supports the limits as given by m = 2. 



The numbers given under m = l are the results of an 

 attempt to apply the same considerations in Ag as were 

 successful for Au * to explain the apparent deviation from 

 the multiple A law in the d 2 (l) sequent. It has generally 

 been supposed that, as in the alkalis, the first D sets of this 

 sub-group depended on m = 2. If so, we have seen that the 

 multiple law is quite definitely broken. To explain this it is 

 assumed that there is a set depending on m = l. If so, the 

 formula shows that the corresponding D lines lie far up in 

 the ultra red. The order of change in the mantissa? is & 

 decrease with decreasing order. The mantissa for m = 2 is 

 close to 35 A 4- 10 5, and if it has to be a multiple of A for 

 m=l, the value should be 35 A or 34 A. 



The mantissa? difference of d(l) and d (2) is always much 

 greater than that of d (2) and <i(3). Here the latter is of 

 the order 14 8, and we should expect therefore the multiple 

 to be less than 35, and probably therefore 34. 



The D n line would depend on ud 1 sequent with a mantissa 

 about 23 Si greater, in line with the observed satellite dis- 

 placement for ??i~ 2. On these bases it is easy to calculate 

 the values of the whole set for m = l and to compare with 

 observational evidence. We have then 



A = 27786-57-l-267f+-l(p 1 -p 3 ), 

 34 A-944743-38-43-078f +3-4(^—^2), 



rf 2 -(l) = N/(l+ 34 A) 2 = 28998-98 + 1-286 f+ ■l{p l -p 2 ).. 



* See the July number, p. 8. 



