CHAPTER VIII. 



AN ADJUSTMENT IN RELATION TO THE FRESNEL COEFFICIENT. 



64. Apparatus, one internal reflection. Figure 86 gives an application of 

 the quadratic interferometer for the possible demonstration of ether drag. 

 Here M, M f , N, N' are the mirrors of the interferometer and L and T the 

 collimator and telescope of this design, the rays passing as shown. In place 

 of the auxiliary mirror to the left of AW, the transparent glass cylinder or 

 disk GG', capable of rotating rapidly around an axis A normal and symmetrical 

 to the ray parallelogram, has been installed. The mantle of the cylinder is 

 supposed to be carefully ground. If the index of refraction GG' is adequate, 

 the component ray reflected at N' takes the path N'C'CNM and then enters 

 the telescope T after reflection. The ray traversing N', on the contrary, takes 

 the path N'NCC'N'M'M and also enters T to interfere with the preceding 

 ray. In other words, the two rays are refracted in like manner by the cylinder 

 GG', at the angles of incidence and refraction i and r, respectively, where 

 t = 2r, since R is the radius of the cylinder. 



f^T~ 



If now the cylinder rotates clockwise to the eye, as shown by the arrows, 

 the rays N'C'CN will be accelerated and the rays NCC'N' retarded by the 

 same amount while in the cylinder, and the question is to what degree such 

 an effect should be observable. 



The adjustment is made for reversed spectrum or linear fringes as presently 

 to be indicated. 



If, in figure 86, the half -silver mirrors, all of equal thickness, are set with 

 their silver reflecting faces as indicated, each ray traverses the glass-path 

 thrice, and since M and N are to be quite equal in thickness, the glass-paths 

 are equal. Not so the air-paths, for these differ by 26. Hence the V-compen- 

 sator VV must be introduced in the ray between N and M, so that the missing 

 26 may be thus inserted, as shown in figure 86. The right-angled mirror V is 

 stationary, while the reentrant mirror V, with its sides respectively parallel 

 to V, is on a micrometer, the screw s' being directed in the direction of b. The 

 displacement is thus virtually normal and the rays CC' are not dislocated on 

 88 



