THE AID OF THE ACHROMATIC FRINGES. 95 



and strong, moreover, although the cylinder used was (as above) an ordinary 

 glass shade. As in case of the triangular interferometer (Chapter X, 82), the 

 fringes rotate when N is displaced parallel to itself or rotated on a vertical 

 axis. To control their size, TV is to be rotated on a horizontal axis. 



69. Modification of the experimental design. Rotating, self=adjusting 

 interferometer. Inasmuch as the experiment (fig. 88) may suffer from inade- 

 quate light, one may notice that there would be no difficulty in rotating the 

 interferometer as indicated in figure 94. Here it is a stout metallic plate or 

 wheel, capable of rotating about the axis a at any reasonable speed. The 

 mirrors N, N f , M, M', which may be as thick as desirable, are rigidly and 

 firmly fastened to the plate. Any displacement from centrifugal force is 

 equally effective in case of both of the rays, and must therefore vanish in the 

 self-compensated interferometer. Any flexure outward of N' would be bal- 

 anced by the outward flexure of m; N would not be flexed and the outward 

 flexure of m' is equally effective on both rays. Hence there should be no 

 appreciable change of size of fringes. Any other stresses due to centrifugal 

 forces could be eliminated by reversing the rotation. The illumination at L 

 is intermittent, but so rapid in succession that a continuous effect is produced 

 to the eye at the telescope T, L and T being adjusted independently of //. 

 The experiment is again favored by the high luminosity of the achromatic 

 fringes. Here, however, it is necessary that an identical glass-path or path 

 of high index of refraction intervenes between each of the mirrors, N, m; 

 m, m'\ m', N'; N', N (fig. 94). For the effect depends on i i/yu 2 ; and as this 

 coefficient may be varied between about 0.3 and 0.7 as extreme limits, it 

 should not go unnoticed. 



