Value of the Silver Oun. 305 



Pass over for the moment the numbers under m = l. 



m = 2. Both D lines can be clearly identified. The D n 

 seems to be split into two of equal intensity, viz. that given 

 in the table (6)42994 and also (6)4300224, which latter is 

 equivalent to a displacement of — $8 1 on the sequent. The 

 latter would give as a limit mean 30646*65. By themselves, 

 therefore, there is nothing to show which is the real Dn, or 

 whether they do not both represent lines displaced from the 

 normal one. The matter would, however, seem to be settled 

 by the D 22 line in which no such ambiguity occurs. This 

 gives a limit mean =31563*28 + , corresponding to D^ (y? ) = 

 30642*85 ± in practical agreement with that from the line 

 42294. The suggested J)u is a 138 displacement on the 

 observed line (2) 43060*02. The numerical agreement is 

 good, but the reality of the indicated connexion has no 

 bearing on the present quest. The maximum observation 

 ■errors taken at d\ = '0b for the short waves may possibly be 

 larger, those for the long waves are inappreciable. The set 

 give reliable material for the determination of £ . 



m = 3. The only directly observed line is D 22 giving a 

 limit mean for D 2 (<x> ) corresponding to D^ = 30640*71 ±1. 

 The suggested D n is the displaced (In) 37518*10 ( - 23 8), but 

 in any case the observational error is so large and the change 

 per oun displacement ( = *72) so small, that it cannot be 

 relied on for an accurate value. The orders above m = 3 do 

 not help in fixing the actual value of f except that they 

 Agree with m = 2, 3 in indicating a negative value of a few 

 units. The following notes have, however, some interest in 

 connexion with the general theory of the analysis of spectra. 



7n = 4. The evidence goes to indicate a greatly diminished 

 mean limit, probably due to the D 2 being both displaced. 

 For the suggested D 22 we find the following linkage 

 •evidence : 



^.(1)39730*79 =35959-79(2), 

 v . (1) 38677*87 = 35961*21 ( 1*5), 

 (1) 33506-45. ti = 35965-09 (*5). 



There appears to be a linked set at b = v= 920*43 ahead, 

 viz. : — 



(32785-17)C4){ (D^-IO.* =32785-10 (-5) 



920-53 *■ ( lw ) 30le8,73 •*= 32785 • 24 (• 5 ) 



(1)29423.88C4) 81564'S (33705-7) j (W7**53.«= 3370853 (*5) 



K} K ' \ (1)31089-11. t-=33705-75 (*5) 



(1)33192*27 =(29421 27) 31563-4 



