In observations to determine the exact positions of celestial objects correction has to be made for the effects of refraction, according to the apparent elevation of these objects in the sky. Such effects are least when the objects in question are directly overhead, for then the rays of light, coming from them to the eye, enter the atmosphere perpendicularly, and not at any slant. A very curious effect, due to refraction, has occasionally been observed during a total eclipse of the moon. To produce an eclipse of this kind, _the earth must, of course, lie directly between the sun and the moon_. Therefore, when we see the shadow creeping over the moon's surface, the sun should actually be well below the horizon. But when a lunar eclipse happens to come on just about sunset, the sun, although really sunk below the horizon, appears still above it through refraction, and the eclipsed moon, situated, of course, exactly opposite to it in the sky, is also lifted up above the horizon by the same cause. Pliny, writing in the first century of the Christian era, describes an eclipse of this kind, and refers to it as a "prodigy." The phenomenon is known as a "horizontal eclipse." It was, no doubt, partly owing to it that the ancients took so long to decide that an eclipse of the moon was really caused by the shadow cast by the earth. Plutarch, indeed, remarks that it was easy enough to understand that a solar eclipse was caused by the interposition of the moon, but that one could not imagine by the interposition _of what body_ the moon itself could be eclipsed. In that apparent movement of the heavens about the earth, which men now know to be caused by the mere rotation of the earth itself, a slight change is observed to be continually taking place. The stars, indeed, are always found to be gradually drawing westward, _i.e._ towards the sun, and losing themselves one after the other in the blaze of his light, only to reappear, however, on the other side of him after a certain lapse of time. This is equivalent to saying that the sun itself seems always creeping slowly _eastward_ in the heaven. The rate at which this appears to take place is such that the sun finds itself back again to its original position, with regard to the starry background, at the end of a year's time. In other words, the sun seems to make a complete tour of the heavens in the course of a year. Here, however, we have another illusion, just as the daily movement of the sky around the earth was an illusion. The truth indeed is, that this apparent movement of the sun eastward among the stars during a year, arises merely from a _continuous displacement of his position_ caused by an actual motion of the earth itself around him in that very time. In a word, it is the earth which really moves around the sun, and not the sun around the earth.