be extinguished. This portion of the image is used in lecture demonstrations as a marked line of reference. But in the lower portion of the image, light has to pass through the air-film, and the bright image suddenly disappears from the screen as soon as the angle of incidence exceeds the critical value.
During a complete rotation of the cylinder, two pairs of readings will be obtained for total reflection. In the first set, one-half of the semi-cylinder will act as the collimating lens, the other half acting as the focussing lens: in the next set (when the cylinder is rotated through 180°) the functions of the two halves are exchanged. From the mean of the differences obtained from the two sets of observations, the value of 2 i is obtained with great accuracy, provided the circle is accurately graduated, and the reading Vernier permits of small angular measurements being made.
The circle used in the simple form of apparatus being only divided into degrees, I could not expect to obtain from it, in the ordinary way, any very accurate results; but by using the following method, it was possible to obtain results which are highly accurate.
The platform index is at first clamped down pointing to the zero of the scale, and the circle as a whole rotated till total reflection takes place. The circle is next clamped, and the platform carrying the cylinder unclamped and rotated in the direction of the lower arrow till light is again totally reflected, the
