304 



Prof. W. M. Hicks on the 





These formulse are exact, but with an uncertainty in f 

 amounting to ±12, they do not afford numerical values of 

 the desired degree o£ accuracy. It is necessary to fix the 

 value of f within limits ±'4. Two directions are available 

 for an attack on this question, (1) the law that the first order 

 d 2 and /sequences depend on mantissse which are multiples 

 of A, and (2) the existence of both difference and summation 

 series. The results of a search for the summation or D lines 

 are given in the following table *, which includes the already 

 known D lines, and the means of the corresponding D and B 

 lines, which give the series limits. These means are printed 

 in italics. Numbers in [ ] are calculated wave numbers,, 

 those in ( ) denote values deduced either from links or dis- 

 placements on observed lines, in a manner described in the 

 text. 



Table of D and D Series. 



28 £] 30644-52 ±1-8 (59643-44) (3-6) 



r 



[1645-62- 



72 

 [1717-62 

 84843 



[2566 '05 



f (6)18270-81 



2025 



m = 2 <{ (10)18291-06 



90032 



(10)19191-38 



f [23731-00] 

 332 



:3 i (8)23734-32 



306 44-52 ±1-8 

 30643-7 2 ±1-5 

 ] 3 1566-32 ±1-5 

 30643- 21 ±-5 

 30642-58 ±'5 

 31563- 28 +-5 



30642-97 ±-7 

 31561 -22±l 

 - (2)26235-29(13-7) 30639-80 ±7 



91711 



' v (6)24651-43(-6) 



:4 -; 91774 



I (1)27153-03(3 



31559-06 + 3-6 



(2)59569-91(3) 



(60566-59) (3) 



(43015-61) (-9) 



(6)42994-11 (-9) 



(2)43935-1909) 



(37551-63)(l-4> 

 (1)38471-01(1-4) 

 (1)35044-43(1-2) 



(35965-09) (-5) 



, (1)27586-13(3-8) 

 <j 918*26 



I (28504-39) (1-6) 



r 



I 



[28404- 19 4- £] 

 92043 



[29324-624-$] 



31560S9 + 1 



(34617-39) (-5) 



30642-94 + U±-3 (32881-7) (-46) 

 31563-05 + ^1 + - 5 (1)33801-48(1-1) 



* Here and throughout figures in ( ) attached on the left of a wave- 

 number denote intensities, when attached- on the right denote possible 

 errors. When the latter are not given by_an observer, these are estimated 

 as -1 for measures to the first decimal, and '05 for measures given to the 

 second. 



