500 



Sir J. Conroy. 



[Mar. 3, 



Red light. Yellow light. Blue light. 



O / 1 O I 



Principal incidence in air observed 76 29 74 37 71 33 

 Calculated from observations made 



in water 77 47 76 30 72 46 



Calculated from observations made 



in carbon tetrachloride . . ? 77 54 77 13 73 52 



In the case of the silver films the difference between the observed 

 and calculated values is greatest with the thinner films (amounting to 

 nearly 4° for the thinnest, with red light) and diminishes as the films 

 increase in thickness. 



The values for the refractive index for gold were also calculated by 

 the formula given by Lundquist (" Pogg. Ann.," vol. clii, p. 405), /w 2 A = 

 tan 2 A(l — sin 2 A. sin 2 2#), and the values for the relative indices for 

 water and carbon bisulphide multiplied by the refractive indices of 

 these media, but the numbers thus obtained did not agree with those 

 deduced from observations of the principal incidence and azimuth in 

 air. 



These experiments appear to show, first, that with glass in contact 

 with media other than air, the tangent of the angle of polarisation is 

 equal to the relative refractive index of the media. 



Second, that the optical constants of a polished metallic plate 

 depend, to a certain extent, on the substance with which it has been 

 polished, and that this surface condition is a fairly permanent one, 

 not being destroyed by contact with liquids or by a considerable 

 amount of rubbing with a clean chamois leather. 



Third, that when the plate is in contact with media other than air, 

 the principal indices are lower, and the principal azimuths higher than 

 in air ; but that there is no simple relationship discoverable between 

 the change in the values of the constants and the indices of the media. 



Fourth, that with metallic films the principal incidence and the prin- 

 cipal azimuth increase with the thickness of the film, and that therefore 

 more than one layer of molecules is concerned in the act of reflexion. 



The fact that with the increase of the thickness of the films the 

 principal azimuth increases appears to show that the surface layers 

 reflect light polarised in the plane of incidence more abundantly than 

 light polarised perpendicularly to that plane, and that as the number 

 of reflecting layers increases, the amount of light polarised perpendicu- 

 larly to the plane of incidence which is reflected increases also ; or, in 

 other words, the light polarised perpendicularly to the plane of inci- 

 dence penetrates to a greater depth than that polarised in the plane of 

 incidence. 



