DE. W. M. HICKS 5 A CEITICAL STUDY OF SPECTRAL SERIES. 429 



In No. 1 the sounded line F 2 is on the verge of the observed region. In No. 2 the 

 sounded is about 4 too large, and there appears a S, displacement on the last. The 

 u, v links themselves are too short to reach. 



m = 3. [37650-6 from Calculated F, 37717 from Observed (5J,) F.] 

 (3) 23973-69.t> (0) 15944 (W.).2v (2) 18739 (C.R).2t; + 4p 



37654'29 + cfc 5651+dv 43305'2 + 2^ 2795 + 4p 46100'2 



Here the reproduced line refers to the normal line calculated from the formula. 

 The v link is too short for the second and third lines, 2v reaches it, but 2v on the 

 first would require a reference line in the ultra-red. Also 15944 is possibly 

 (-7^(2). 



m = 4. [34779'80.] 



(3) 23584 (C.R.).~) 



> (-2<S,)(3) 16725-1 (W.). 



(<?,) (0) 23598-2 (W>,J 



34776-0 5655'2 + de-du 4043 1 '2 



Again note a C.R. copper line supported by a W. and R.R. displaced <\ line 



m = 5. [33259-17.] 



(2) 252137 (W.).r (()) 18024 (W.).e 



33259-05 5635'25 + dv 38894'3 2808'0 + de-dv 41702'3 



Here appears the frequent 5635. It is 5649 14, that is, there is a limit 

 displacement of ^ in the second line and an extra one in the third, making the 

 second separation 2816. 



m = 6. 



The same links cannot serve as sounders for all three lines. 



For F, is (2) 211725'4 (W.).u = 32364'34 + dw as against calculated 3236214. 

 For F 2 is v.(2) 16816'5 (W.}.e.u = 3800^8 +de + du-dv. 

 For F,, (4) 18448 (C.R.).2w = 40831' 



These give separations 5640'5, 2826'8, where the sum has the normal value. In 

 (-^) (3) 23973-69 (W.).w = 35178'45 which is 2816'2 ahead of P, we have the 

 completion of a mesh with the other three lines. 



The foregoing discussion has shown : (l) that this F set belongs to the 1864 type 

 discovered in X ; (2) that the usual displacements are present and that the calculated 

 value per oun 13"96 satisfies all the numerical relations formed; and (3) that the 

 lines observed with copper electrodes by CAMERON and RAMSAY seem to belong 

 specially to parallel series to this set. We could feel complete confidence in the 

 allocation of the lines were it not that the a constant in the formula is positive, and 

 that the line m = 6 is not reproduced more closely. The absence of direct 



