



of Light at Certain Metal-Liquid Surfaces. 89 



n' no Observer 



Mercury in HC1 (dilute) 1-44 1-39 Des Coudres 



" " " (cone.) 1-30 1-39 " 



" Na,S,0 8 (cone.) 1-32 1-42 " 



" " " (sup. sat.)-- 1*28 1-43 " 



" " petroleum.. 1'38 1*44 " 



" " chloroform 1-37 1-44 " 



" olive oil.. ... 1-47 1-47 " 



" " turpentine 1"25 1*47 " 



" " " 1-32 1-47 " 



" CS 2 1-59 1-63 



" " " 1-48 1*63 " 



Now it is impossible to ascribe these discrepancies in every 

 case to films of surface contamination ; for they occur for those 

 metals (notably copper and mercury) for which entirely satis- 

 factory results in air have been obtained, and under circum- 

 stances apparently precluding the formation of any chemical 

 film. Hence, as we have seen that it is highly probable that 

 a perfectly clean metal possesses no appreciable transition layer 

 when in contact with air, it is only possible to ascribe these 

 discrepancies to a transition layer due to the liquid. Now an 

 inspection of the above table shows that the index calculated 

 from the reflection from the metal is, in all but four cases, less 

 than that deduced by the direct refraction methods. On the 

 theory as developed by Drude,* this result must be interpreted 

 as due to a film or layer having an index less than that of the 

 liquid. This suggests that the effect is due to a film of air or 

 other gas caught between the liquid and the mirror, since the 

 transition layer of a liquid-air surface would have an index less 

 than that of the liquid. It is, moreover, natural to expect the 

 presence of such films due either to entrainment in the act of 

 immersion or to dissolved gas in the liquid, more probably the 

 former. 



On the other hand, if it is assumed that no such gaseous 

 film is present, the observed discrepancies must be attributed 

 to a transition layer due to the liquid. In view of the fact that 

 the effects of the transition layer in the case of transparent 

 solids in air are inappreciable, it would seem that the layer in 

 question here cannot be a region of interpenetration of the 

 metal and the liquid. Consequently we must conclude (if we 

 reject the possibility of the gaseous film) that the observed dis- 

 crepancies are due to a layer at the surface of the liquid which 

 is independent of the medium with which it may be in con- 

 tact and which has optical properties different from those in 

 the interior. According to the theory, the index of such a 

 layer will be less than that of the interior if n' is less than n a ; 



*Wied. Ann., xxxix, p. 539, 1891. See also, below, under heading, 

 " Theory." 



