THE AID OF THE ACHROMATIC FRINGES. Ill 



displacing B parallel to itself. The slit-images may be superposed by rotating 

 A on a vertical axis, and the local coincidence is then secured at B as specified. 

 The fringes should then be in the field. From the complete compensation 

 they are liable to be large. Rotation of B or C on a vertical axis moves the 

 slit-images through the field of the telescope. Horizontal axes (preferably 

 at A) provide for vertical coincidence. Reflection from the glass face can 

 usually be blotted out by a small screen on C, particularly when s is reversed. 



It seemed reasonable to suppose that the rays of this instrument could also 

 be appreciably separated, for instance, by moving B to B', as the rays b, c 

 would interfere in the telescope. But this attempt did not succeed to a useful 

 extent. In fact, even slight displacement of the mirror B passes the fringes 

 through a maximum with rotation. The same thing happens on rotation of 

 any mirror, on either a horizontal or vertical axis, so that the fringes are large 

 and fugitive and difficult to control. To obtain considerable displacement 

 of B, the mirror C must be reciprocally rotated on a vertical axis with the 

 fringes continually in the field. Similarly B and C may be rotated together, 

 etc. The apparatus is nevertheless interesting, and if optic plate were used 

 (any wedge angle in A is here of serious consequence) and small angles of 

 incidence were chosen, it might be useful. 



Thus, for instance, I had hoped to use the apparatus for an experiment on 

 the Fresnel coefficient (see Chapter VIII) for the case of two internal reflec- 

 tions in the rotating cylinder G, figure 89. The cylinder is in such a case 

 obviously used to greater advantage. But apart from the general difficulty 

 of this installation, the low intensity of the twice-reflected spectra militated 

 against the practical availability of the method, which is unfortunate, as the 

 two spectra (unlike the case above) have the same focus in T. There being 

 respectively two and three reflections, the fringes are those of reversed spectra. 



To put fringes of any particular type in the white field, the mirror B, figure 

 109, may be rotated on a horizontal axis. If a vertical group is produced in 

 this way, this may be enlarged indefinitely by displacing B parallel to itself. 

 The fringes then remain vertical throughout, but their direction of motion 

 changes (with abrupt rotation) at the maximum. If B is rotated on a vertical 

 axis, the fringes in the now moving slit-image similarly change size, remaining 

 vertical. The phenomenon is here not apt to be symmetric on the two sides 

 of the maximum. Inclined fringes, while following the same scheme of enlarge- 

 ment, continually change their inclination, the fringe pattern being ultimately 

 hyperbolic. 



83. Herschel's fringes. Herschel's fringes, as produced by the familiar 

 apparatus consisting of a right-angled prism reposing with its broad face on 

 a plate of obsidian, present the well-known group of achromatic fringes run- 

 ning parallel to the arc or limit of total reflection. Observation is made in a 

 direction normal to the edge of the prism. 



It occurred to me that the phenomenon could be made much more striking 

 and of wider scope if a long 60 prism were used and observation made in a 



