598 



Stoney — Cause of Double Lines in Spectra. 



The numbers in brackets are determinations not made by Kayser and Runge. 

 They are the earlier determinations reduced to Rowland's scale. 



An accurate determination of the values of /i throughout the spectrum is very 

 much wanted. However, so far as AN is concerned, /x differs so little from unity 

 that AN need not be distinguished from Ak, until much more refined observations 

 are made than those hitherto recorded. 



These three optic notes P, Z>, and S, with possibly a fourth one (the existence 

 of which there is some slender ground to suspect), make up the optic chord emitted 

 by the molecules of sodium. Nor is the case of sodium an isolated one : all the 

 other light monad elements, Li, K, Rb, Cs, emit optic chords of essentially the same 

 character, consisting of three notes P, D, and S, closely resembling those of sodium 

 in many important respects. 



The optical clang of these notes — the relation in which their partials stand to 

 one another — may be roughly exhibited to the eye by plotting them down on 

 separate maps of oscillation-frequencies, when the pattern which the lines make 

 becomes conspicuous. The scale of the diagram is too small to show that any of 

 the lines are double. In fact, most of them are so (see foregoing Tables). 



A = T//. 



S S 



400 



300 



Ultra-violet. 



Visible part of the spectrum. 



Ultra-red. 



The spectrum as seen has the much more disorderly appearance which would result 

 from plotting them all down on one map. 



Professor Rydberg is of opinion that the value of Ak (the interval between the 

 constituents of each pair) is the same in all the pairs of sodium, and that the 

 recorded discrepancies are due to the roughness of the observations. This is a 

 matter that careful observations will decide. Meanwhile we are concerned with 

 studying the inferences that can be drawn; and in order to do this it will be 

 convenient to take series S first. 



