Page:Collected Physical Papers.djvu/43

Rh commence the experiment with an angle of incidence greater than the critical angle, the incident ray being then totally reflected. The angle of incidence is next slowly decreased till its value is slightly less than the critical. At this point the ray is refracted into the air, making an angle of 90° with the normal to the surface; a receiver fixed against the side of the semi-cylinder at R, responds to the refracted radiation. A slight increase of the angle of incidence produces total reflection and the receiver ceases to respond.

The platform on which the cylinder rests, carries the usual index. Rotation of the cylinder in a left-handed direction gives a reading for total reflection. A second reading is obtained by rotation of the cylinder in the opposite direction, the receiver being now at R′. The difference between the first and the second reading is evidently equal to twice the critical angle.

A metallic screen with a small central opening is placed against the plane face of the semi-cylinder for utilising only the central rays. In order that all the rays should undergo total reflection simultaneously, it is necessary that the rays incident on the plane of separation of the two media should be parallel. This is effected by the cylinder itself. From the approximate value of $$\mu$$ found from a preliminary experiment, the focal distance of the semi-cylinder is roughly calculated. The spark-gap of the radiator is placed at this focus, and the rays thus rendered nearly parallel. Each subsequent experiment gives a more accurate value of $$\mu$$, and from the corrected value of the focal distance thus obtained, a more accurate adjustment is made for the next experiment.