Maintenance oj Periodic Motion by Solid Friction. 161 



normally had no electric vector swinging in this direction, 

 and so the author measured only the normal effect which 

 increases continuously with decreasing wave-length. 



(5) Elster and Geitel investigated the influence of polari- 

 zation only in that part of the spectrum in which the K-Na 

 alloy has its selective effect. Consequently they found only 

 for E L, that at different angles of incidence the emission of 

 electrons was proportional to the absorbed energy ; and the 

 ratio of the effects in the two main planes of polarization 

 had naturally values, which like 1: 50* are much too large 

 to be only explained by the difference in the absorption of 

 light, and which prove the existence of some special efficacy 

 of Ell [19]. 



(6) The same was true in Pohl's experiments, in which 

 above the ultraviolet end u of the selective effect (fig. 2) 

 the singularity of E || disappears, so that only the normal 

 effect is left. 



Inasmuch as all the contradictions which we spoke of in 

 the beginning are explained now, we think that we are 

 justified in saying that a new selective effect exists, apart 

 from the normal one. We do not as yet know much about 

 the true nature of this effect, but from experiments made on 

 K-Hg alloys [37] and on the influence of the angle of 

 incidence [27], we conclude that it is a molecular resonance 

 phenomenon, in which the electrons follow directly the 

 electric vector ; at any rate it cannot be explained by the 

 " Quantentheorie '■ of Planck-Einstein. 



We are continuing the researches with the aid of a grant 

 from the Jagor Fund (Berlin), for which we wish to express 

 our best thanks. 



Physikalisches Institut der Universitat Berlin, 

 October 1910. 



XIX. On the Maintenance of Periodic Motion by Solid 

 Friction. By Andkew Stephenson f . 



1. nPHE maintenance of periodic motion by solid friction 

 _L demonstrates that such friction diminishes as the 

 velocity increases through a small range at least. However 

 the friction may vary there is always a position of equili- 

 brium and the small motion about it is evidently of type 



x+ (fc — X)x + n 2 x = Q, 



* We got even the ratio 1 : 300 in K-Na alloys, in which the normal 

 effect was comparatively much smaller than in fig. 3. 

 t Communicated by the Author. 



Phil Mag. S. 6. Vol. 21. No. 121. Jan. 1911. M 



