of the Metals. 171 



of the spectrum with precision * Inasmuch as all the rays passed 

 through the glass plate bearing the metal, they would all be re- 

 tarded thereby in a similar manner. Indeed the above-described 

 arrangement of the experiment even allowed the displacement of 

 the interference-bands due to the metallic films to be estimated 

 independently of the errors of the plane-parallel glass plates, so 

 that one might even employ a piece of very good common mirror- 

 glass. 



Two other plane-parallel glass plates were also placed in the 

 path of the interfering bundles of rays ; these were fastened to 

 the horizontal axis of a goniometer, in such a way that their 

 refracting surfaces formed a small angle with one another, and 

 cut one another in lines parallel to the axis of the goniometer, 

 perpendicular to the direction of the interfering rays. 



The ray I. went through the one, the ray II. through 

 the other of these plane-parallel glass plates, which were thus 

 differently inclined to the two rays. The paths traversed by the 

 two rays through the plates were equal in one particular position 

 of the axis of the goniometer ; on turning the axis towards the 

 right, the thickness of one of the plates increased faster than 

 that of the other ; on turning it towards the left, the opposite took 

 place. Thus the pair of plates formed a compensator of Soleil 

 and Duboscq's construction*, such as Jamin has already em- 

 ployed in experiments with an interference-apparatus. 



One is able at will, by means of this compensator, to retard the 

 ray I. or II. to a slight extent, and so to destroy the displacement 

 of the bands of interference in the upper relatively to those in the 

 lower part of the spectrum, a displacement produced by the dif- 

 ference between the velocity of the rays passing through metal 

 and through ait in the upper part of the spectrum. If, in order 

 to neutralize this displacement, it were necessary to retard the 

 ray passing through metal, then the velocity during the passage 

 through metal would be the greater ; if the ray passing through 

 air required retarding, then the velocity in air would be the 

 greater. 



In the whole of these experiments the transparent metal plate 

 was placed nearly perpendicular to the interfering rays, in order 

 that a displacement of the interference-bands due to the elliptical 

 polarization of the light passing through the metal might be 

 avoided. 



Silver, which was transparent and had a blue or violet colour, 

 or gold, transparent and of a brown or blue-green colour, effected 

 a displacement which was destroyed by a retardation of the ray 



* See Jamin, Ann. de Chim. et de Phys. (3) vol.xlix. (185?) p. 288; and 

 vol.lii. (1858) p. 166. 



