Now this 'systematic motion,' which would render all calculations as to the sun's motion inaccurate or even altogether worthless, is by many astronomers held to be an observed reality. The star-drift, first pointed out by Proctor, has been shown to exist in many other groups of stars, while the curious arrangements of stars all over the heavens in straight lines, or regular curves, or spirals, strongly suggests a wide extension of the same kind of relation. But even more extensive systematic movements have been observed or suggested by astronomers. Sir D. Gill, by an extensive research, believes that he has found indications of a rotation of the brighter fixed stars as a whole in regard to the fainter fixed stars as a whole. Mr. Maxwell Hall has also found indications of a movement of a large group of stars, including our sun, around a common centre, situated in a direction towards Epsilon Andromedæ, and at a distance of about 490 years of light-travel. These last two motions are not yet established; but they seem to prove two important facts--(_a_) that eminent astronomers believe that _some_ systematic motions must exist among the stars, or they would not devote so much labour to the search for them; and (_b_) that extensive systematic motions of some kind do exist, or even these results would not have been obtained. Mr. W.W. Campbell, of the Lick Observatory, thus remarks on the uncertainty of determinations of the sun's motions: 'The motion of the solar system is a purely relative quantity. It refers to specified groups of stars. The results for various groups may differ widely, and all be correct. It would be easy to select a group of stars with reference to which the solar motion would be reversed 180° from the values assigned above' (_Astrophysical Journal_, vol. xiii. p. 87. 1901). It must be remembered that, within a uniform cluster of stars, each moving round the common centre of gravity of the whole cluster, Kepler's laws do not prevail, the law being that the angular velocities are all identical, so that the more distant stars move faster than those nearer the centre, subject to modifications, however, due to the varying density of the cluster. But if the cluster is nearly globular, there must be stars moving round the centre in every plane, and this would lead to apparent motions in many directions as viewed by us, although those which were moving in the same plane as ourselves would, when compared with remote stars outside the cluster, appear to be all moving in the same direction and at the same rate, forming, in fact, one of those drifting systems of stars already referred to. Again, if in the process of formation of our cluster, smaller aggregations already having a rotatory motion were drawn into it, this might lead to their revolving in an opposite direction to those which were formed from the original nebula, thus increasing the diversities of apparent motion.