444 DR. W. M. HICKS: A CRITICAL STUDY OF SPECTRAL SERIES. 



This gives for m = 3, 4, 5, the lines n = 44834'16, 47323'04, 48672'56. The O-C 

 for the first is dX = - '06 whilst the others should be weak and are near to the limits 

 of observations. The links are so small that as sounders they produce lines also close 

 to the observed region. For m = 1, the difference of the sequent mantissse of 23899 

 and 23782 (D 12 ) is 4144 and 4A 2 -6<S, = 71^= 4141'2, or exact with dX = '013 on 

 two lines. It will be shown later (p. 447) that D 12 depends on the A 2 multiple. The 

 series is, therefore, probably a real doublet series with this for the outer satellites. 

 For m = 2 the D 13 set with a displacement of about 17lf(5 is inserted since it recalls 

 the order of magnitude in Kr and X, whilst the two D 12 satellites for m = 2 and 3 are 

 inserted because they show the same displacement of 39^<l An extrapolated line on 

 an observed one [2361T5], 179'5, (2) 2379TO gives a displacement 175cS for m = I 

 corresponding to that for m = 2. It may also be noted that in Argon the third lines 

 of the triplets appear to be missing in this D series. 



The following arrangement would seem to indicate the existence of a set parallel to 

 the above : 



The [ ] in the second column are calculated from the first column by adding 20"65. 

 The line in ( ) is determined as the mean of the linked lines u.(l) 47784"02 = ... 44'55 

 and (l) 46183"62.*M.e = ... 42'80. As 5<5 on the limit produces a separation of 20"65, 

 the lines in the second column would correspond to the parallel displaced series 

 ( 5<J)D n . The separations of the last column increase with the order to a normal 

 value of i/! and point to a constant displacement in the sequent. 



With the above D set should be associated a doublet F series whose F : ( oo) = d u (l) 

 = 51731-03-23899'OS = 27831'95 and separation that of the D satellite or 116'5. In 

 searching for this allowance will have to be made for the prevalence of displacement 

 in the lower orders. As a fact, however, this only appears for m 3 in F,, although in 

 F 2 there are considerable signs of it. The following table with succeeding notes 

 contains the data, including also those for the corresponding F series ; also for certain 

 lines which show a separation about 150, and which may probably be the analogue of 

 the 1864 X series. 



We shall denote this last set by dashing the letters. 



The formula obtained from the two first F! lines and the given limit is 



-95-N/{ 

 I [ 



n = 27831-95-Nm + 



m 



