LIGHT AND ELECTRIFICATION. 185 



of being perfectly smooth and horizontal ; it thus makes 

 easy definite optical experiments on angles of incidence and 

 azimuth of polarisation. Employing the potassium-sodium 

 alloy in a hydrogen-vacuum, Elster and Geitel found that 

 the most effective light is that which is polarised perpendic- 

 ularly to the plane of incidence, i.e., light whose vibrations 

 have a component normal to the surface. Light vibrations 

 parallel to the surface are comparatively inoperative. 



It this be so, of course it follows that with common 

 light normal incidence should be less effective than 

 oblique incidence ; and indeed it may be supposed 

 that any effect of normally incident rays might largely 

 be due to the scratched and unpolished state of 

 ordinary metallic surfaces. Experiments conducted in my 

 laboratory on polished surfaces of silver do not bear this 

 out. Normally incident light has a very considerable 

 effect, at least on ordinary surfaces in air, though obliquely 

 incident light has a greater. Using the liquid K — Na alloy 

 it was possible to verify completely Elster and Geitel's 

 result with respect to the azimuth of polarisation, and we 

 also found that with common liQfht and different angles of 

 incidence the effect increased rapidly as the incidence 

 changed from normal (o°) to something near 6o°, but that 

 it then ceased to increase any more. So far as I have 

 observed, the effect of polarisation and varying incidence 

 is much more marked with these alkaline metals in vacuum, 

 subjected to long waves, than it is with such surfaces as 

 polished zinc or silver in ^ordinary air and other gases, 

 exposed to ultra-violet waves ; though with these also it is 

 perfectly real and definite. 



I have now rehearsed what may be considered as a 

 summary of most of the more prominent and interesting 

 facts connected with the subject, abstaining from any 

 theoretical comment. In a future article the attempt may 

 be made to collate these facts, and to consider their bearing 

 with reference to a possible explanation or perception of 

 the meaning of the effect, so far as it is reasonable to sup- 

 pose it understood at present. 



Oliver J. Lodge. 



