420 



DK. W. M. HICKS: A CEITICAL STUDY OF SPECTEAL SERIES. 



Taking this as F ( oo) and using the /(m) sequences as given by the formula (p. 412) 

 it is possible to calculate the positions of the lines in question. The results are given 

 down to m = 8. 



. S l displacement on F ( <x) gives a change 6 -045.] 



2. P requires sounding. (2) 35923 -SO.e.v = 47371 '08. 



3. (-5Sj) (3) 28063-32 = ...93-54; F requires sounding. (8,) (1) 37508-25.* = 41930-21. 



4. (-28!) (2) 39419-27 = ...31 '36. 



5. (- 28,) (2w) 31934-91 =....47-00; (28,) (1) 38089' 14= ...77-05. 



6. For F. (In) 37263-13 ; also (-68,) (<1) 37226-34 = ...62-88, or the same. 



7. v.(l) 37720 -05 = ...92'05; also (- 38,) (4) 33274- 11 = ...92-24; (2) 32304-23.?; = 36732-23. 



The mean limit is 3501 2'37. The set form an additional test that the extrapolated 

 lines 16013 really exist. It is curious to note that the even orders of F only show 

 directly observed lines, whilst the odd show displaced limit lines. The F lines are far 

 to the violet end and come into the observed region only when weakened by high 

 order. The F lines are all linked to lines of higher frequency by the 1864 link, also for 

 m = 4, 6 to lines of lower frequency. The same tendency is shown in F to lower 

 frequency, any such to higher frequency lead to unobserved regions. This fact is 

 important as showing that at least here the 1864 separation enters in the link 

 relation, and not as a direct displacement on the limit. 



To the F 3 (o) limit corresponds a 1864 triplet series parallel to that originally 

 considered. I have been able to follow it up in the same way as the foregoing as far 

 as m = 26 at least. It accentuates the evidence for the displaced sets but as that is 

 sufficiently supported by the results already discussed it would seem unnecessary to 

 overburden the present communication with additional detail. Whenever the actual 

 relations of the various displaced lines to one another are the subject of discussion 

 these details will be of the first importance. A knowledge of these relations should 

 be expected to throw a flood of light on the constitution of spectra, but this new 

 question cannot be taken up here. It may be noted, however, that the first F and F 

 lines of the triplets up to m = 10 are given in the table under F 3 . 



