Stoney — Cause of Double Lines in Spectra. 599 



Series S. 



8 (a). If, as appears to be the case, Series S consists of double lines none of 

 which has a satellite midway between its components, it follows from our investi- 

 gation that — 



The path of the electron, from which this series arises, is aplane curve. 



8 {h). If further, as Professor Rydberg supposes, the A/c's (or rather the AiV's, 

 which are practically undistinguishable from the Ak's) are identical in the several 

 double lines of the series, it will follow that — 



The dominant orbit of the electron (as started by the last encounter 

 with another molecule) is affected during the subsequent flight of the mole- 

 cule by an apsidal perturbation, which carries the orbit as a tvhole round 

 in its own plane, without altering its form. 



S (c). If, moreover, as Professor Rydberg concludes from the observations, 

 the less refrangible line of each of the pairs is the brighter, it will follow that — 



The elliptic partials of which the undisturbed orbit consists are all de- 

 scribed in the same direction, and that this direction is the reverse of that 

 in which the apsidal motion takes place. 



8{d). No satellites such as those described in Problem III., p. 576 have been re- 

 corded in connexion with the lines of this series. If fui'ther observation establishes 

 the fact that none such exist, it will follow that — 



The perturbating forces do not occasion any precession. 



8 (e). Professor Rydberg concludes from the observations that Ak for this 

 series = "146. The correction which should be applied to this to allow for the 

 dispersion of the air is inappreciable, so that we may take 0-146 as the value of 

 AN, the difference of the oscillation-frequencies of the two rays. Now, by Problem 

 II., n (the frequency of apsidal circuits) = ^ AJV. It is therefore 0-073 or nearly 

 1/14. If this detei'mination is correct it follows that — 



One apsidal circuit lasts during 14 jots of time; so that about 30 of these 

 will on the average be described during each flight of the molecule 

 between its encounters. 



8 (/). Meanwhile a vast number of revolutions in the elliptic orbits of the 

 partials will have taken place, ranging from 1226 during each of the 30 apsidal 

 circuits in the case of the least refrangible of the observed double lines of the 



