DOUBLE STARS About the year 165o Jean Baptiste Riccioli made the first dis covery of a double star, viz., Mizar or Ursae Majoris, the middle star in the tail of the Great Bear. Looking at Mizar with the naked eye, we see a fainter star, Alcor, very close to it ; it is a fair test of eyesight to separate them. Their separation is how ever, 1', and although there is no doubt that Mizar and Alcor have a physical association, since they are found to be moving at just the same rate across the sky, the term double star is only applied in case of much closer proximity. Riccioli's discovery was that Mizar itself consists of two stars separated by 15" or about of the distance between Mizar and Alcor. This would nowadays be regarded as a wide pair, and many of the most inter esting double stars have a separation under I". The actual dis tance between the two components of Mizar is not less than II times the distance of Neptune from the sun. The two stars must be attracting one another, but the gravitational control across such a distance is so feeble that they will take more than 3,00o years to complete their orbital revolution. It is, therefore, not sur prising that the measures made during the last 15o years fail to indicate definitely any orbital motion; and the only evidence of genuine connection (apart from the improbability of two inde pendent bright stars lying so closely in the same line of sight) is afforded by their common motion across the sky, and common radial velocity.
Similar discoveries followed rapidly. 0 Orionis (in the midst of the great nebula) was resolved by Huyghens in 1656; 'y Arietis by Hooke in 1664. In the southern hemisphere the duplicity of a Crucis and a Centauri was discovered by Jesuit missionaries in 1685-89. The foundation of systematic measurement and study of double stars is due principally to Sir William Herschel and Wilhelm Struve. The latter, working at Dorpat between 1813 and 1835, formed a catalogue of 3,110 pairs as the result of sur veying about 120,000 stars.
The number of recognized visual double stars is now about 15,00o, but the number is necessarily somewhat arbitrary since, if too wide limits of separation are admitted, stars which are only in accidental proximity may be included. Among the naked eye stars, one in nine is a telescopic double. It is certain that the true proportion of binary systems is much greater than this, many being too close to be resolved with the telescope. Thus, taking the nearest known stars (which give the best chance of observing duplicity), eight out of 20 are double.
Catalogues of double stars no doubt contain a number of spurious pairs, i.e., stars which are not actually near one another in space and merely happen to lie in the same line of sight, but considerations of probability show that there will not be many of these. The physical connection is generally shown, as in the case of Mizar, by common proper motion. If spectroscopic deter minations of radial velocity have been made these will provide an additional check. The connection implied is evidently the con nection of a common origin. The two stars were born from the same nebula; they originally shared in common the velocity of the parent nebula, and they have gone on moving together be cause they have been subjected to the same influences ever since. In most of them the gravitational tie is too weak to resist any differential disturbance, e.g., another star passing between the pair. The fact that so many of these wide doubles survive teaches us that chance approaches of stars to one another must be very rare.
Of greater interest are those double stars close enough together to show the effects of their mutual gravitation. There are about 120 double stars for which it has been possible to work out defi nite orbits. In about 700 more there is a rough indication of orbital motion, but the arc described since observation commenced is too short to determine the orbit satisfactorily. The shortest periods are 5.7 years (6 Equulei) and 6.9 years (13 Ceti) ; about 6o pairs have periods of under 1 oo years. It is natural to enquire whether the planes of the orbits tend to lie in any particular direction—parallel to the galactic plane, for example. A curious difficulty arises in this investigation. If the reader has watched the revolution of a cup-anemometer (or of a garden-sprinkler) he will probably recall that the direction of revolution appears obvious and unmistakable, but after blinking the eyes the revolu tion seems just as unmistakably in the opposite direction and in a different plane. For the same reason we always find two planes of rotation of the double star, either of which may be the true one. We can only adopt both planes, and investigate as best we can the resulting mass of data, half of which is correct and half spurious. So far as can be ascertained the planes of double star orbits are quite at random.