454 DE. W. M. HICKS : A CKITICAL STUDY OF SPECTRAL SEEIES. 



taking the first two lines as really in the relation indicated, and are thus enabled to 

 arrive at the value of F (6). Using B.M.M. measures the separations given by the 

 F (2) lines are their differences from 17081*46 or 94'88, 323'55 with very small errors 

 '05. These, therefore, give for P (6) 34336'06-25-94'88 - 3422118'25 and 

 33918'08"5 + 323'55 = 34221'63. Both are therefore within error limits of their mean 

 34221 '40. With this for F (6), and the limit corrected as above to 31851 '13 the 

 value of F (6) is 29460'96. With the value calculated from the formula O- C = - '23. 

 As illustrating the way in which D and F sequents are subject to displacements, it 

 is interesting to notice that although the separations 94, 323 do not appear as 

 directly dependent 011 F lines after m = 2, they nevertheless occur in the 

 neighbourhood. It will be sufficient here only to refer in detail to the case of 

 F (4) = 27009-47. At about 50 ahead of this there is a line (6) 27060'60 (W.). With 

 this there is the following scheme : 



2G76G'3(L.D.) 96'6 26863'4 (L.D.) 197'2 (0) 27060'60 (W.) 321'4 27382 (L.D.) 



As the L.D. lines are subject, even if correct to the nearest A.U., to equally probable 

 errors between d\ = '5, or here to dn = 4, these separations correspond to the 

 94'88, and 323'55, whilst 197"2 is the link e. The displacement of 27060 from F (4) 

 may then be in the F sequent or the limit probably the former, or a, b or c link. 



The existence of the parallel sets, the indications of the associated F types, and 

 the nature of the displacements all point to the conclusion that our original series is 

 of the F type. Consequently the limit F ( o) = 3185T13 is a d ( I ) sequent, but 

 there is nothing as yet to show whether it belongs to a d n or a satellite set. If, 

 however, we can find a D (l) set the value of D ( co) can be obtained with sufficient 

 accuracy to obtain the values of A,, A 2 and the e link. The further consideration of the 

 F series will therefore be postponed until this further information has been obtained. 



It has already been remarked that only the S (2) and D (l) lines in any S or D 

 group can be expected within the observed region. A superficial inspection of the 

 list of lines shows a very large number of separations in the neighbourhood of 46 

 to 49 and about 20, clearly related to i/j, v 2 values as they show the normal ratio v l : v 2 , 

 and are in step with those of the other gases. In about the region in which the 

 D lines should be expected BALY gives the strong set (4) 21200'90 49'24 (4)2125014 

 with an equally strong line at (4) 21230'! 1 which might be a D n line to the doublet 

 D J2 set. WATSON, however, gives other strong lines as well, including a triplet. 

 These give 



(1) 2115677 (W.) 48-18 (l) 21204'95 (W.) 19'71 (4) 21224'66 (B., W.) 

 44-13 45-19 



(4) 21200-90 (B.) 49-24 (4) 21250'14 (B.) 



(5) -81 (W.) 49-36 (5) '19 (W.) 

 29-21 



(5) 21230-11 (B., W.) 



