594 Stoney — Cause of Double Lines in Spectra. 



may be completely determined by observation of the line. Now it is probable that 

 all the lines of the spectra of H, Li, Na, K, Rb, and Cs fall into one or other of these 

 categories, and if this is so, all the partials of the cethereal undulation causing these 

 spectra can he determined by observations tvith the spectroscope. 



We have next to consider the motions going on in one of the molecules of the 

 gas. These correspond to the motions going on in an orchestra ; and of them the 

 only ones to which we have a direct clue are the motions of the electrons them- 

 selves, since we can only work back from the effect on the sether and these are 

 the only motions in the molecules that act directly on the sether. They correspond 

 to the motions of the sounding-boards of the stringed instruments and the columns 

 of air in the wind instruments of the orchestra, as these are what act directly on 

 the surrounding atmosphere. 



Here, again (see last chapter), the motion of the electron may be resolved into 

 elliptic partials of the form, 



X = a cos 6t, y = h sin 6t ; 



but as the electron is not necessarily confined to one plane, we must remember that 

 its path may be a cu.rve of double curvature, in which case the elliptic partials 

 will lie in different planes, see p. 591. 



CHAPTER VI. 



APPLICATIONS. 



We may, however, best form a judgment as to how far observations with the 

 spectroscope will carry us by studying some one spectrum. The spectrum of any 

 of the above-mentioned monads would do, but amongst them that of sodium seems 

 the best adapted for our purpose. 



Professor Rydberg has collected and analyzed the principal observations on 

 this spectrum, and I follow the selection from amongst them which he made, 

 adding to them the determinations since made by Professors Kayser and Runge. 

 Thirty-five lines,' thirteen double lines and nine others, had been observed.* Pro- 

 fessor Rydberg conjectures that they are all, except one satellite, double. In this 

 connexion it should be noted that, in accordance with the analysis given above on 

 page 572, even where there is the apsidal motion which produces double lines, one 

 of a pair may be of cypher intensity, and, accordingly, the line as it appears in 



* Two others more refrangible, and belonging to series P, have since been announced by Professors 

 Kayser and Runge. 



