Binary stars were incidentally discovered by Sir William Herschel. In his researches to get a stellar parallax he had selected a number of double stars for test purposes, on the assumption that, if one of such a pair were much nearer than the other, it might show a displacement with regard to its neighbour as a direct consequence of the earth's orbital movement around the sun. He, however, failed entirely to obtain any parallaxes, the triumph in this being, as we have seen, reserved for Bessel. But in some of the double stars which he had selected, he found certain alterations in the relative positions of the bodies, which plainly were not a consequence of the earth's motion, but showed rather that there was an actual circling movement of the bodies themselves under their mutual attractions. It is to be noted that the existence of such connected pairs had been foretold as probable by the Rev. John Michell, who lived a short time before Herschel. The researches into binary systems--both those which can be seen with the eye and those which can be observed by means of the spectroscope, ought to impress upon us very forcibly the wide sway of the law of gravitation. Of star clusters about 100 are known, and such systems often contain several thousand stars. They usually cover an area of sky somewhat smaller than the moon appears to fill. In most clusters the stars are very faint, and, as a rule, are between the twelfth and sixteenth magnitudes. It is difficult to say whether these are actually small bodies, or whether their faintness is due merely to their great distance from us, since they are much too far off to show any appreciable parallactic displacement. Mr. Gore, however, thinks there is good evidence to show that the stars in clusters are really close, and that the clusters themselves fill a comparatively small space. One of the finest examples of a cluster is the great globular one, in the constellation of Hercules, discovered by Halley in 1714. It contains over 5000 stars, and upon a clear, dark night is visible to the naked eye as a patch of light. In the telescope, however, it is a wonderful object. There are also fine clusters in the constellations of Auriga, Pegasus, and Canes Venatici. In the southern heavens there are some magnificent examples of globular clusters. This hemisphere seems, indeed, to be richer in such objects than the northern. For instance, there is a great one in the constellation of the Centaur, containing some 6000 stars (see Plate XXI., p. 306). [Illustration: PLATE XXI. THE GREAT GLOBULAR CLUSTER IN THE SOUTHERN CONSTELLATION OF CENTAURUS From a photograph taken at the Cape Observatory, on May 24th, 1903. Time of exposure, 1 hour. (Page 306)]