Stoney — Cause of Double Lines in Spectra. 603 



From these values it appears that the partial which causes the great yellow 

 double line of sodium is one in which the electron makes m/n = 1984 elliptic 

 revolutions, while the apsidal motion carries the ellipse once round. An apsidal cir- 

 cuit is completed in 1/w = 1 1 '7 jots. And there is time for about 420 22 = 36 of these 

 complete apsidal circuits to take place during the average flight of a molecule be- 

 tween two consecutive encounters, assuming this journey to occupy about 420 jots. 



The more refrangible of the two lines is known to be the brighter ; and I hope 

 soon to have the means of making a good determination of the ratio of their 

 brightnesses. Meanwhile the best estimate I can at present make gives this ratio 

 as lying somewhere between 3 : 2 and 4:3. Now these numbers are nearly in the 

 ratios of 36 : 25 and 49 : 36. We may assume, therefore, that [a -f ly : (a — bf lies 

 somewhere between these ratios, and that, therefore, a : h lies somewhere between 

 11:1 and 13:1. Accordingly the partial in this case is a long-shaped elliiDse, in 

 form somewhere between the two ellipses delineated in the figure. Round 



Fig. 13. 



this ellipse the electron travels 1984 times while the ellipse shifts gradually once 

 round in the same direction, and something like 36 of these slow apsidal circuits 

 are performed dm'ing each rectilinear flight of the molecule. These are the 

 events that occur in the partial which gives rise to the great yellow double line of 

 sodium ; and an equal amount of information may he obtained in the case of every other 

 double line that can be observed tvith the requisite accuracy. 



It has been mentioned that the three series P, Z>, and S ajDpear in the spectra 

 of all the light monad elements, except hydrogen, viz. of Li, Na, K, Rb, and Cs ; 

 and it may be added that they are found in positions in the spectrum of pro- 

 gressively lower refrangibility in the order in which the elements are named, 

 i.e. in the order of their atomic weights. Another remark that should be made is, 

 that when we compare the spectra of these elements with one another the value 

 of Ak is found to increase with the atomic weight, showing that the apsidal motion 

 is swifter, and, therefore, that the perturbating force is stronger in the more massive 

 molecules. A series, very much like one of the foregoing, is found in the spectrum 

 of hydrogen, but it is in a situation of too low refrangibility to be any one of 

 these in the case of an element with such low atomic weight.* 



* One is almost tempted to conjecture that all the light monads, including hydrogen, have similar 

 spectra, and that there are four series in each, H, P, D, and S, of which JI appears in the spectrum of 



TEANS. BOT. DFB. SOC, N.S. TOL. IT., PART. XI. 4 Q 



