( ^'i ) 



observes for some values'). That Ritz has not thought of the character 

 of summational and difTerential vibrations appears from the fact that 

 no attention is paid in his paper to the /?z/'t'?«5/(y of tlie lines which occur 

 in the combinations. Rrrz thinks that in the observations of the new 

 lines he tinds a support for the view that the spectral formuhie 

 must be written with two order numbers. It is, however, evident 

 that every spectral formula which correctly rei)resents the fundamental 

 lines, also holds for these new lines, so that the combination prin- 

 ciple must not be considered as an advantage of Ritz's spectral 

 formula. It is better to account for these lines without using a specti-al 

 formula. That Ritz thinks of something else than summational and 

 dilferential vil)rations appears further in his paper: "Das magnetische 

 Feld im Atom kann bei allen Spektren betrachtet weerden als erzeugt 

 durcli zwei Pole entgegengesetzten Vorzeichens; jedes der zwei Glieder 

 von {m,a,(i) — {n, a! , (i') stellt je den Einflusz eine.s Poles dar; jeder 

 dieser Pole kann im Atom verschiedene Lagen annehmen, die z.B. 

 bei Wasserstoff iiquidistant auf einer Geraden liegen. Vertauschung 

 dieser Lagen untereinander entspricht dem Kombinationsprinzip. All- 

 gemeiner kann man wohl die Vermutung aussprechen, dasz sick die 

 einfacheu Gesetze auf die Lagen dieser Pole ini Atom beziehen.'' ^) 



I have found that it is the intensest lines of a spectrum (especially 

 the first doublet of a principal series, which give a flame its charac- 

 teristic colour) that give rise to the existence of combinational 

 vibrations. The intensity of the summational and differential vibrations 

 is, however, comparatively slight; uncertainty may prevail with 

 regard to the lines lying in the infra-red ; for a number of lines in 

 the visible spectrum the slight intensity appears already from the 

 fact that they have not been seen by earlier observers. 



The presence of combination lines points to the existence in the 



1) Phys. Zeitsclir. 16, 1908, p. 523, -'bei den oben sub 1, 3, 4, angefiihrtcn 

 Fallen hat man sogar aussclilieszlich Diiïerenzen ocler Sutnmen beobachteter Wellen- 

 zahlen zu berechnen." Tlie combination i' ^=^ ("^5 Pi ^) — {i^i pi ^)» '^ ^ ^i '^i '^i ••• 

 e.g. is to be considered as v = {(l,5,s,rj) — (m,;?, jr)} — {(1, 5,s, (j) — (2,/',rr)| 

 loc. cit. p. 521. However also the cases 6 and 7 p. 52:2 must be considered as 

 summational and differential vibrations: 



v=^{2,p,^)—{m.p,-p,,:T,-.^,)=\{2,p,rr)—{34,(f)\ + \{^A(i)—^^^ 

 loc. cit. p. 521 and 522. 



— 1(3, (/, Ö) — (4, p,-p^, ^i-^^Jl loc. cit. p. 522. 

 The signification of case 5 is further elucidated in this paper. 



2) Loc. cit. p. 523 and 524. Also W. Ritz, Magnetische Atomfelder u. Seriën - 

 spektren, Ann. d. Physik 25, G60, 1908. 



32* 



