Series System in the Spectrum of Gold. 15 



For direct evidence we must look, as before, for F lines in 

 the ultra violet. There are no lines at the places indicated. 

 The nearest are (6)53401-69 + 2-9p, (-70, 183), and 

 (2)53l51-91 + 2-9 ( — 174, 32) with the proper separation of 

 249*78 +. If they are the sought F lines the complete set 

 would be : 



F : -181-9 + x 26485-0 53151'91+2'9, 



F 2 66 + .v 26733-9 53401-69+2-9, 



with x of order +4, depending on obs. errors 2'9p and limit 

 uncertainty. These are 6Q± less than the calculated values 

 and can be explained by a +7 8 l displacement in the sequent, 

 which produces a separation of 64*75. The u, v links are too 

 small to reach into the observed region. We have no 

 knowledge as yet as to how links should be attached ,to wave- 

 numbers which formulae give as negative, but we should 

 expect doublet separations to be conserved. We should then 

 look for lines, with added and with subtracted links as 

 follows : — 



F : e-181-9 + ^=17071-8 + ^', -^-181-9 = 17435-7 +ar, 

 F 2 e + 66+x 1 =17319-8 + ^', —e + 66 =17187-8+^. 



Now we find in this neighbourhood a large number of 

 lines with all the signs of being debris, so to say, of a 

 shattered system, viz., weak lines of the same nature and 

 separated by multiples of the same numbers. Taking the 

 lines as they occur in order they may be arranged as follows : — 



e.~F 2 . e.Fj. 



(2 w) 17410-02 

 ■(1;/-) 17178-50 



9-3 29-57 = 6x4-93 (4) 17675-53 are. 



[17187-8 +.«'] [174357 4-#] [17634-6] 



251'8-A" (2») 17439-59 25151 O) 17691-10 



4 99 4 92 



(In) 17444-58 251-44 (1») 17696-02 



419 

 11-48 (It) 17700-21 



689 



(3)17456-06 251-04 (3r) 1770710 



439 



(3») 17711-49 



911 

 (3) 17720-69 



