14 Prof. W. M. Hicks on the 



errors and are shown as alterations on the observed mea- 

 sures : — 



D. D- 



( 2735-94 29469-85 [56203-77] 



381556 \ 24888 248 ' 88 



I 2984-82 29469-85 55953-45 + 1-44 (p= -o) 



[ 6551-30 33285-41 60021-04-4-68 (p = r3) 



This would give d 2 (l) = 29469*85-2735-94 = 26733-91. 

 It may be taken as this with an uncertainty at present of 

 ±4. This makes A = 111940, 3 = 1406-78 br within the 

 limits of the direct determination from S(x>). 



The F and ¥ series. These series have as their limits the 

 first Dn, D12 sequents. With our new D(l) lines or the ch 

 sequent with a mantissa 9 A. 



F a (»)=d a (l) = 26733-9--237* a? not > 10. 



The D satellite displacement of 27 8i gives a separation of 

 24889, or 28 S ± of 258*05. We should expect, therefore, 

 doublet series with these separations on both sides of 26734, 

 appertaining to F 1 (oo) = 26486 or 26476*81. We find on 

 inspection a very large number indicating a profusion of 

 displacements, and the lines in consequence so crowded that 

 it is difficult to determine definitely whether numerical co- 

 incidences correspond to real relationships or not. For our 

 present purposes we desire to obtain the F 2 (l) lines. Now, 

 like the d sequences the/(l) sequence also has its source in 

 a mantissa which is a multiple of A and at the same time is 

 a large fraction approximating to unity. Here, then, the 

 multiple can only be 9 or 8, or possibly 7 or 6 — or there may 

 be, as in the rare gases, several groups depending on both. 

 In other elements we find indications of: F satellites depending 

 only on small displacements of order 2B ± . Also the lower 

 orders are often subject to very large A displacements. We 

 proceed to discuss the two cases of 9 A and 8 A in order : — 



Case I. /(I) = N/(9 A) 2 . In this case d,(l) and/(l) are 

 identical and equal to 26733'96+2. The expected lines 

 should therefore be : 



F 2 (1) = F 2 (l)=53467-8±4 3 



F 1 (1)= -248-89 F 1 (l) = 53218-9dz4. 



F 2 (1)=0 would correspond to no oscillation and probably 

 to instability, and lead to an expectation of displacement. 



