CHAPTER III. 



THE REFRACTION OF LONG GLASS COLUMNS MEASURED BY 

 DISPLACEMENT INTERFEROMETRY. 



38. Introductory. The measurement of indices of refraction and their 

 differences for different colors, in terms of the shift of the ellipses in the spec- 

 trum, seemed to give an opportunity for unusual sensitiveness of method 

 when long columns of glass are inserted in one of the interfering beams. But 

 this expectation was not realized in full, as the amount of shifting per unit of 

 displacement of the micrometer mirror decreases with the thickness of the 

 glass, or the length of the column. The measurements made are nevertheless 

 interesting as a test of the availability of the equation 



., , A dp 



N e e-n cos R 5 -rr 



cos R dA 



where N c is the codrdinate of the movable mirror, corresponding to the center 

 of the ellipse at the wave-length X of the spectrum. R is the angle of refrac- 

 tion, p. the index of refraction of the glass for the same color (the angle of 

 incidence at the grating being I), and e the thickness of the glass column in 

 the direction of the penetrating ray. For most purposes the Cauchy equation 

 may be used for determining dn/d\. 



39. Glass columns. The column to be tested is placed with its faces normal 

 and symmetrical to one of the component beams, the corresponding mirror 

 having been advanced proportionately to the length of the column, until the 

 ellipses appear. The fine adjustment is then made with the micrometer screws, 

 the displacement A7V needed to shift the centers of the ellipses from one spec- 

 trum line to the next in succession being observed. This is the chief datum 

 of interest in the present paper; for from it the tests in question may be 

 constructed. 



The first columns provided were made from thick glass rods, the ends of 

 which had been ground off normally to the axis of the cylinder by Mr. Petit- 

 didier. But in none of the rods prepared was the glass sufficiently homogene- 

 ous to admit of its use. In some cases, in fact, the stress or its equivalent 

 was so strong as to make the rod virtually opaque and the polarization figures 

 correspondingly intense. The attempt was made to remove the stress by 

 annealing at low red heat, but this also was quite unsuccessful. 



Failing in all other trials, I finally resolved to build up columns from ordi- 

 nary plates of glass, cemented together with Canada balsam. In this way I 

 72 



