DE. W. M. HICKS : A CRITICAL STUDY OF SPECTRAL SERIES. 459 



characteristic of D linkages. As referring to the F series we find, starting from 20551 



which corresponds to 31851 



(3) 20556'84 

 44-13 

 (0)20551-35 49-62 (2) 20600'97 



(0) ... 2'02 85-81 



114-84 (0) 20642-61 



(5) 20666-19 

 157-51 



D n (3) 20708-86 

 194-42 



(2) 2074577 



232-02 



(3) 20783-37 



249-22 

 (1) 20800-57 



269-39 312-66 



[2082074] [48-78] (2) 20869'52 



275-00 (0) ... 72'56 }> 5'61 



2-57 J 



(1) 20826-35 48-78 (4) ... 7513 



There is also (l) 20542'57 at 166'29 behind the D n line 20708'86 which seems related 

 to Rossi's series referred to later. 



The application of the D qualification test is unfortunately here "nugatory on 

 account of the smallness of the oun. It is, however, striking that so many of the F 

 separations put themselves in evidence. This agreement adds considerable weight to 

 the allocation. The evidence is also practically convincing when the WATSON 

 separations are brought into evidence as is done a few paragraphs below. Support is 

 also given in the doublet system shown in connection with 20826, which show the 

 same small displacements as are exhibited in the first of the parallel F sets adduced 

 above. In this case to 20869 as a D 2 would correspond a [20820] as a Dj 269"39 

 ahead of 20551, whereas the F is 31273 behind the corresponding F! line (17081). 

 If the two sets correspond, i.e., the d sequent and the F ( oo ) for this set are the same, 

 the D group must belong not to 20551 but to a parallel D group in which the limit 

 is 31273 269"39 = 43'34 behind the normal D(o). Now this is the separation of 

 20556 behind 20600. In fact the separation of 20556 and 20869 is 312'66. This 

 further indicates that 20556 is a D 2 line in the position shown in the above table, in 

 which the corresponding D : line is too weak to have been seen. This also gives another 

 F separation 85 with 20642. The change 44'13 on the D limit requires a displace- 

 ment on the mantissa of 611 and 42^ = 611 '3, which is exact. The v l corresponding 



