Stoney — Cause of Double Lines in Spectra. 601 



Finally the lines are diffuse. If it be further the case that they are similarly 

 diffuse, and that the thicknesses of the lines (wlien plotted down on a map of 

 oscillation-frequencies) are everywhere the same, we must attribute the diffuseness 

 to a common cause, wbich may be that the apsidal motion of the dominant orbit 

 of electron D is not a shifting of the orbit with uniform angular motion, as it is in 

 the case of electron >S', but that there is a subsidiary perturbation of this motion 

 which bears to it the same relation that nutation does to precession in the rotation 

 of the earth. The oscillation-frequency of this nutation can be ' determined by 

 measuring the thickness of the lines. See p. 581. As the lines are found to be more 

 winged on their less refrangible side, the subsidiary perturbation that causes them 

 is more complex than a mere pendulous oscillation ; but until all that observation 

 can tell us is known it would be useless to search further for the cause. 



I have not taken into account the two lines which are close to the third term of 

 series D, and which Professor Rydberg regards as satellites to that term. Professors 

 Kayser and Runge conjecture that they do not belong to any of the series P, D 

 and S, but that they are the only visible term of a fourth series of which the rest 

 lies beyond the parts of the spectrum that have been explored ; and they point out 

 in support of their view that the constituents of this pair are winged towards the 

 violet, while all the lines that are known to belong to series D are winged towards 

 the red. Their positions forbid our attributing them to the circumstances which 

 may produce quadruple lines sketched out in Problems V. and VI. However, 

 we shall be in a better position to deal with this group of four lines when more is 

 known of their distances and intensities. 



Series P. 



The remaining lines of sodium that have been observed, including the great 

 yellow double line, form another natural group which Professor Rydberg 

 calls the principal series. All but one of the terms of this series are of high 

 refrangibility. Some of them are known to be double lines; in others only a 

 single line is (as yet) recorded. If they turn out to be single, they probably arise 

 from partials that are circular. The recorded observations upon the spectra of 

 sodium, potassium, and rubidium show that A/c is not the same in all the pairs of 

 series P, which indicates that the perturbating forces acting on Electron P 

 are such as to induce different rates of apsidal shift upon the several partials of 

 its dominant motion. Accordingly, the dominant orbit undergoes a change of form 

 as well as of size and position during the flight of the molecule. The double lines 

 of series P are characterized by having their more refrangible constituent the 

 brighter, in which respect they differ from the double lines of series D and >S'. It 



