Page:Popular Astronomy - Airy - 1881.djvu/181

Rh practically is inapplicable, because the uncertainty of refraction is great, for the air is in a heated state, and it is therefore in a state unfavourable for the observation; and because no star can be observed near the sun. In this case, a small uncertainty produces a far greater effect than in the case of the moon. Suppose, for instance, that there is an uncertainty in the observation which will produce an uncertainty of one second in the horizontal parallax. The smaller the parallax is, the greater is the distance in the same proportion. If, then, by this error, the moon's horizontal parallax is altered from 57 minutes to 57 minutes 1 second, the measure of distance is altered by only $1⁄3491$ part of the whole, or by 70 miles. But if the sun's horizontal parallax is altered from 9 seconds to 10 seconds; the measure of distance is altered by $1⁄10$ part of the whole, or by more than 9 millions of miles. On this account, the method of simple parallax entirely fails in ascertaining the distance of the sun.

I then mentioned another very ingenious method, one founded on the observation of the place of the moon when it is "dichotomized." This is a Greek expression used to denote that state of the moon when it is half illuminated. If we can fix upon that time exactly, we shall know that the angle at the moon made by lines drawn to the sun and the earth is a right angle, and if we can then measure the angle at the earth between the sun and the moon, and subtract that angle from 90 degrees, we shall have the angle at the sun made by lines drawn to the earth and moon; and from this we shall be able to compute the proportion between the sun's distance and the moon's distance. The method fails because the surface of the moon is so exceedingly rough.