45 G DE. W. M. HICKS : A CEITICAL STUDY OF SPECTRAL SERIES. 



and triplet lines in question. The points against the allocation of these lines to the 

 normal D group are, (l) the satellite separations are not in the usual ratio 5 : 3 (they 

 are in fact close to the ratio 3 : 5), (2) there is no appearance of satellite lines 

 corresponding to the parallel F lines noted above, and (3) the limit 53081 is somewhat 

 larger than we should expect from the march of the limits in the other gases. 



Taking it, however, as the limit, it is possible with the given value of Aj to 

 calculate the link e. The result is 197 + 4d]. The occurrency curve as found from 

 B.'s and W.'s observations is given in Plate 2, tig. 6. The dotted line is the result 

 when the numerous rough measures by L.D. are included. As is seen the maxima 

 occur at 195 '6 and 198, the calculated, at a minimum. The peaks look as if analogous 

 to corresponding peaks found in other elements, but the analogy is doubtful. In the 

 other elements these (much larger ones) are produced by the prevalence of displace- 

 ments by a few ouns, and by Aj operating on (xS) S ( oo ). Such changes here would 

 be very small and the corresponding effects are really shown by the flattened tops of 

 the peaks. The peaks themselves are due in all probability to another cause to be 

 considered shortly (p. 458). 



In the region in which the S triplets should be expected are found (all interferential 



measures) 



(6) 28787-86 49-81 (5) 28837'67 18-61 (4) 28856'28 



with intensities in the proper order, although the v lt v a are slightly different from 

 those obtained in the D set. The interferential measures of MEGGERS and of B.M.M. 

 differ considerably ('008) and are not so reliable therefore as usual. BALY and 

 WATSON agree in giving the separations as 4972, 18 '64. The third line, however, 

 has already appeared in the quite definite relation 1429 + 417 ahead of F (4). It 

 cannot be S (2), but the latter may be a weak line close to it. 

 There is also another doublet set 



(10) 27818-89 47-75 (l) 278G6'64\ 



49-93 (1) G8'82J 



(8) "73 48-07 (1) 6'80 



The suggested S 2 line is abnormally weak. There are, however, here a 

 number of other weak lines which have the appearance of displaced debris. If the 

 first set form a S group, the second would belong to a group with smaller limit, 

 which is also indicated by the smaller triplet separations. The limit 5308 1'24 was 

 obtained from apparently the only stable D set, whilst the first set are apparently 

 the only stable strong S group. It is natural therefore to take its limit as the same. 

 If so, the limit for the second, being 969'03 (or '35 dv) less, is 52112'20. The two 

 mantissas are 437420-13'54 and 450723-13'92^+13'92^. Their difference is 



13303-'38+ 13'92^ = 20(665'15--019f+7cZv) = 



