Fortunately, we have such a means of calculation, and it has been employed very extensively at different periods, chiefly by M. Biot in France, and quite recently by German astronomers, in calculating the positions of the stars from the present time to a period of 2000 years B.C. We can thus determine with a very high degree of accuracy the latitude, longitude, right ascension, declination, and the relation of the stars to an equinox, a solstice, or a pole, as far back as we choose. Since we have the planes of the equator and ecliptic cutting each other at different points in consequence of the cause which I have pointed out--the attraction of the sun and moon--we have a fixed equator and a variable equator depending upon that. In consequence of the attraction of the planets upon the earth, the plane of the ecliptic itself is not fixed, so that we have not only a variable equator, but also a variable ecliptic. What has been done in these calculations is to determine the relations and the results of these variations. The calculations undertaken for the special purposes of this book will be referred to later. A simpler, though not so accurate a method consists in the use of a processional globe. In this we have two fixed points at the part of the globe representing the poles of the heavens, on which the globe may be rotated; when this is done the stars move absolutely without any reference to the earth or to the plane of the equator, but purely with reference to the ecliptic. We have, then, this globe quite independent of the earth's axis, flow can we make it dependent upon the earth's axis? We have two brass circles at a distance of 23½° from each pole of the heavens (north and south); these represent the circle described by the pole of the earth in the period of 25,000 years. In these circles are forty-eight holes in which I can fix two additional clamping screws, and rotate the globe with respect to them by throwing out of gear the two points which produced the ecliptic revolution.