to the fact, that the stars move all in a piece—either that they are connected with some one thing turning upon an axis, or that they stand still while the earth turns round an axis of its own; one or other of these things is certain.
Having now come to that result, as one which is generally established, I shall just mention a slight departure from it. Perhaps you may be surprised to hear me say the rule is established as true, and yet there is a departure from it. This is the way we go on in science, as in everything else; we have to make out that something is true; then we find out under certain circumstances that it is not quite true and then we have to consider and find out how the departure can be explained. Now this is the fact. When we have a telescope of considerable power attached to the Equatoreal, so that we can see a small departure from the centre of the telescope in the position of the star we are looking at, and when we trace the course of that star down to the horizon, we find this as the universal fact that though the instrument be set up as carefully as possible, yet the star is not quite so near the horizon as we are led to expect. What can the cause be? There is a consideration that explains it perfectly—it is what is called refraction.
In order to see what refraction is, we may advantageously examine refraction on a larger scale. In a room generally darkened, let a lamp be introduced, as at A, Figure 3, and let it shine through a hole B in a screen CD, so as to produce a spot of light E on the wall. Place in the course of that ray of light a trough F, whose sides are pieces of plate glass. Now pour some water into the trough, and see what effect it produces. You will observe that the light is
