﻿and the Theory of Atomic Structure. 



553 



lines 7] and ff. 2 were usually measured only by reference to 

 other lines this table is not very well adapted for an accurate 

 calculation of the energy difference between the N 4 and N : > 

 level (see diagram IV.). The lines 79 and /3 l4 are dealt with 

 in § 11 of Part IV. The last column gives the frequency 

 difference between the absorption discontinuities L x and L 2 

 as measured by the author. 



Table XVII. gives those differences of the square roots of 

 the frequencies, which according to Part II. §5 are nearly 

 constant. As the absolute errors in the frequencies are 

 about inversely proportional to the second power of the 

 corresponding wave-length, it is scarcely surprising that 

 there are more irregularities in this table than in the corre- 

 sponding Table V. 



§ 5. In Table XVIII. the L absorption discontinuities of 

 some of the elements are given. They were found on the 

 same plates where the emission-lines were photographed 



Table XVIII. 

 Absorption discontinuities. 





Wave-lengths. 



Frequencies. 



L, L, 



1 ' : 



L, L 2 



58 Ce 



2159-7 1 2005 

 (2158) ! (2007) 



421-94 



(422-3) 



454-44 

 (454-0) 





59 Pr 



2072-7 1919-7 



43944 j 474-68 

 (440-0) (474-1) 





(2071) 



(1922) 



60 Nd 



1990-3 



(1992) 



1837-6 | 



(1842) 



457-86 ; 495-90 



(457-5) (494-7) 

 49502 





62 Sm 



1840-9 

 1773 



... 



1 



63 Eu 



51407 





(compare Part II. § 9), and were measured relatively to /3 2 

 or 7 t of the same element. Hertz's values are added in 

 parentheses ; they agree better with ours for the elements 

 Ce, Pr, and Nd than for Ba and Cs * (see Tables XI. and XII. 

 Part II. § 9). Absorption-discontinuities may be measured 

 very easily in this w T ay, but the method is only applicable if 



* It may be stated that my values for the latter elements agree very 

 well with those of Mr. Lindsay (compare Part III. § 8). 



