458 
DR. W. M. HICKS: A CRITICAL STUDY OF SPECTRAL SERIES. 
limit and that the present pair run parallel to one 26(1 above the old If this existed 
it should be 21142'G4, 21191'98, 21211'69. There are lines for the last two but none 
for the first. They are 
[21142-74] [49*24] (0) 21190-61 20*10 (6) 21210-71. 
There is thus considerable evidence for the existence of parallel S groups as well as 
of parallel D ones. Each set has its corresponding separations, but with the 
same oun multiples for Ai, A.j. We should consequently expect to find the presence 
of corresponding a ... e links. With the S limit 52112 the e link is found to be 191-49. 
Now the occurrency curve for e gives a maximum between 195-6 and 196 pointing 
to a S limit as basis about 650 less than 53081 or, say, 52430. This makes e — 195-63, 
)'i = 43 - 9 . Do we find evidence for S and D sets about this amount less in wave- 
number than the old ? For the S we are landed in a region which forms a gap in B.’s, 
or W.’s o])servations but which contains a number of lines by L.D. Amongst these 
we find 
Si. 
S 2 . 
1429 1 
195-7 
^27964 + 4 
47 
28011±4 
105*9 J 
1 
213 
28177F4 
40 
28217±4 
The suggested Sj, Sg fall into the proper positions, the link 1429 is one of Watson’s 
constants, 195-7 is the e link proper to the limit, 105-9 is u or v, and 213 is 2 m or 2v. 
For the D sets we find amongst W.’s lines 
( 0 ) 20551-35 49*62 (2) 20600-97 
157*51 
(3) 20708-86 
The 20551 is 601 below D^g = 21156, which when the variability of the satellites is 
considered may be taken as the analogue of 610 for the S set. If it is corrected* by 
dn = -7 {d\ — —-17) the becomes the correct value 48-9, and it is then also separated 
from 20708 as a D^ line by 156-8, which at once suggests the origin of the 156 
parallel F and F sets previously brought to light. If this relation be real, the old 
F ( 00 ) 31851 is a d satellite sequent, belonging to 20551-35 or ...2-0 and the limit is 
20552-01-31851-13 = 52403-13. Not only is the 156 separation found, but there are a 
large number of lines in the neighbourhood which give, possibly within the error limits, 
the other separations indicated by the traces of parallel F series adduced at the 
beginning. Also as indicating a D region we find large repetitions of e and h links 
■’'‘Not necessarily error, probably the usual D displacement on sequent. 
