112 



DISPLACEMENT INTERFEROMETRY BY 



plane of symmetry parallel to the edge of the prism. In the interest of variety, 

 moreover, it is preferable not to employ strictly accurate surfaces, so that the 

 prisms with which grandfather used to decorate his gas-fixtures will, as a rule, 

 suffice admirably. In figure no, P is such a prism (truncated) on a plate of 

 obsidian Q, the long edges being normal to a white window curtain at L near 

 by, illuminated with sunlight or daylight ; or any light toward the front over- 

 head is good. 



The rays that are wanted, s, will enter symmetrically at a mean angle of 

 about 30 to the vertical and after reflection at the base of the prism and the 

 plate reach in the eye in the direction E. The rays totally reflected, t, come 

 from a greater angle to the vertical and are not wanted. 



--s 



The limit of total reflection here (also easily recognized) is usually a sharp 

 parabolic or cuspidal apex. The light seen through either face enters by the 

 opposed face. On looking down from a steeper angle and with properly 

 selected faces, brilliant groups of complete confocal ellipses (major axes 0.5 

 to over 2 inches), or of confocal hyperbolae may be seen in each of the roof- 

 faces. To find advantageous face combinations, the three faces of each prism 

 should be examined in succession, and it is well to rub P on Q to improve the 

 contact. On moving the eye fore and aft or using different pressures, any 

 type of ellipse with white or colored disk may be produced at pleasure. It is 

 usually preferable to use a shorter plate Q than is given in the figure, about 

 one- half the length of the prism. 



When well produced the ellipses may also be seen by side-light, with different 

 patterns in the two roof-faces. 



The type of interference figure clearly depends on micrometric differences 

 of the faces in contact. The ellipses are Newton's rings modified by the color 

 dispersion of the glass. The hyperbolae, however, are about equally frequent ; 

 but their character is less easily stated. They probably originated in cylindrics. 



84. The same, continued. The case of the 45 to 90 prism (fig. in) with 

 the right-angles faces, respectively horizontal (on the plate) and vertical, is 



