446 



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



Notes. Displacements are in evidence except for FI. A desired wave-length can be obtained very 

 closely since a S t displacement produces a change of only 42. Consequently in deductions by displace- 

 ment the results have little weight, and in fact would have none at all were it not that we now know as a 

 fact that such exist. Failing such likely displacements, recourse has been had in a few cases to linked 

 lines and here their evidence has weight. 



m = 2. For FI, (-16) (1)42192-13 = 18-21; (98) (1) 42233- 12 = 18-45; mean taken. For F 2 

 (-58)(1) 42326-07 = 34-22. 



m = 3. There is no observed F,, but (3) 19351 -20.c = 20070-91 or with E.V. 71' 65 ; 



FI (3) (7A 2 ) = 20103-85 and 20105-52 is observed by E.V. d\ = - 42. 



ForF 2 , (-385!) (4)20174-83 (E.V.) = 90-22; (20^) (2) 20198-39 = 90-29 mean taken. 



ForFj, (-58)(1) 35580-73 = 88-88, also F 2 , (-5S)(1) 35695-03 E.V. 2 = 703 -18 2. 



m = 4. The observed line for FI is much stronger than should be expected ; also its - C is in the 

 opposite direction to that of the others. It may be displaced by A 2 on the sequent and so intensified. 

 The A 2 would produce - C = "00 and the observed = Fj (4) (A 2 ), or it may hide the real FI. FI 

 is inserted as 32819- 150-50. It only indicates that 32819 is the correct value of F'. F 2 is given as 

 (58) (1) 32793-11. 



m = 5. F! shows a link e = 719-81 to (3) 25241-75. F 2 is entered as (188) (4) 24665-59 = 37-25, 

 but has no weight. FI = (15S) (1) 31279-10. 



m = 6. F 2 is the mean of (- GSj) (2) 25536-15 = 45-87 and (78) (1) 25557-12 = 45'78. The 

 calculated value of FI gives a better 151-7 to F'i. 



m = 7. The calculated FI is 26009 30. The deduced value is (1) 25290 92.e = 26010 73, which again 

 has an apparent approximate e link = 720 -77 forward to an E.V. line at (2)26731-40. This is 

 however a coincidence as the last line is S 2 (2). In connection with the linked line 26010-73 maybe 

 taken the pair of lines (-68) (3) 26001-11 = 10-89 and (7J8) (3) 26022-85 = 10 '63 whose mean agrees 

 precisely with the former. For F 2 a similar split with two may be observed ( - 2|8)(3) 26123'03 = 27-07 

 and (28) (1)26130-61 = 27-30. 



m = 8. F 2 is (5) 25796'57.e ; also ( - 28) (2) 26512-75 = 16-05. It may be noticed that the separa- 

 tions to F! and F' are both about 2- 3 too small, or 1|S on the limit. F! is (2) 28540 -26.0 and F 2 is 

 (1) 28659 'O8.e. These linked lines in this order are therefore reliable. 



m = 9. F 2 is (1) 28389-59.e. 



It should especially be noted how with increasing order the normal calculated lines 

 appear as observed with good intensities. As in the other instances it suggests itself 

 that this is due to the diffusion of the energy of the lower order lines into the 

 formation of numerous displaced ones. The usual accurate determination of the limit 

 as the value of (Fj + Fj) is not here applicable as, for the reason given above, the 

 determination of the actual displaced lines is unreliable. These values are printed in 

 italics. The more reliable results point to a limit higher than that calculated in 

 S ( co ), with about + 2. 



