immediately thrown to the top of the wall, as at G. If the hole in the screen be so large that it is not
Fig. 3.
entirely covered by the trough, there will still remain a little light on the wall below, which shows the original direction. You will now see how much the direction of the light has been diverted by the action of the water in the trough. That effect is produced by the refraction of the water. It did not exist before the water was there, but it does exist now that the water is in the trough. I will now show the bearing of this matter on the subject of the disturbance in the position of stars. Figure 4 represents the prism of water we have been looking at.
Fig. 4.
The effect of it is this—a beam of light, coming in the direction of the line AB, does not pursue its original direction, but when it is received by the