the complement of the whole light of the fainter star finds the retina more sensible than that part which is identical in colour with the brighter star, and the impression of the complementary colour accordingly prevails. In many cases, however, the difference of colour of the two stars is real.

เ

When the colours are complementary, the more brilliant star is generally of a bright red or orange colour, the smaller appearing bluish or greenish. The double stars Cancri and y Andromeda are examples of this. According to Sir J. Herschel, insulated stars of a red colour, some almost blood-red, occur in many parts of the heavens; but no example has been met with of a decidedly green or blue star unassociated with a much brighter companion.

699. Triple and other multiple stars. When telescopes of the greatest efficiency are directed upon some stars, which to more ordinary instruments appear only double, they prove to consist of three or more stars. In some cases one of the two companions only is double, so that the entire combination is triple. In others both are double, the whole being, therefore, a quadruple star. An example of this latter class is presented by the star Lyræ. Sometimes the third star is much smaller than the principal ones, for example, in the cases of Cancri, & Scorpii, 11 Monocerotis, and 12 Lyncis. In others, as in e Orionis, the four component stars are all conspicuous.

700. Attempts to discover the stellar parallax by double stars.-When the attention of astronomers was first attracted to double stars, it was thought they would afford a most promising means of determining the annual parallax, and thereby discovering the distance of the stars. If we suppose the two individuals composing a double star, being situate very nearly in the same direction as seen from the earth, to be at very different distances, it might be expected that their apparent relative position would vary at different seasons of the year, by reason of the change of position of the earth.

Let A and B, fig. 99, represent the two individuals composing a double star. Let c and D represent two positions of the earth in its annual orbit, separated by an interval of half a year, and placed therefore on opposite sides of the sun s. When viewed from c, the star B will be to the left of the star A; and when viewed from D, it will be to the right of it. During the intermediate six months the relative change of position would gradually be effected, and

Fig. 99.

the one star would thus appear either to revolve annually round the other, or would oscillate semi-annually from side to side of the other. The extent of its play compared with the diameter C D of the earth's orbit, would supply the data necessary to determine the proportion which the distance of the stars would bear to that diameter.

The great problem of the stellar parallax seemed thus to be reduced to the measurement of the small interval between the individuals of double stars; and it happened fortunately, that the micrometers used in astronomical instruments were capable of measuring these minute angles with much greater relative accuracy than could be attained in the observations on greater angular distances. To these advantages were added the absence of all possible errors arising from refraction, errors incidental to the graduation of instruments, from uncertainty of levels and plumb-lines, from all estimations of aberration and precession; in a word, from all effects which, equally affecting both the individual stars observed, could not interfere with the results of the observations, whatever they might be.

701. Observations of Sir W. Herschel.-These considerations raised great hopes among astronomers, that the means were in their hands to resolve finally the great problem of the stellar parallax, and Sir William Herschel accordingly engaged, with all his characteristic ardour and sagacity, in an extensive series of observations on the numerous double stars, for the original discovery of which science was already so deeply indebted to his labours. He had not, however, proceeded far in his researches, when phenomena unfolded themselves before him, indicating a discovery of a much higher order and interest than that of the parallax which he sought. He found that the relative position of the individuals of many of the double stars which he examined were subject to a change, but that the period of this change had no relation to the period of the earth's motion. It is evident that whatever appearances can proceed from the earth's annual motion, must be not only periodic and regular, but must pass annually through the same series of phases, always showing the same phase on each return of the same epoch of the sidereal year. In the changes of position which Sir William Herschel observed in the double stars, no such series of phases presented themselves. Periods, it is true, were soon developed; but these periods were regulated by intervals which neither agreed with each other nor with the earth's annual motion.

702. His discovery of binary stars. Some other explanation of the phenomena must, therefore, be sought for; and the illustrious observer soon arrived at the conclusion, that these apparent changes

E E

of position were due to real motions in the stars themselves; that these stars, in fact, moved in proper orbits in the same manner as the planets moved around the sun. The slowness of the succession of changes which were observed, rendered it necessary to watch their progress for a long period of time before their motions could be certainly or accurately known; and accordingly, although these researches were commenced in 1778, it was not until the year 1803 that the observer had collected data sufficient to justify any positive conclusion respecting their orbital motions. In that and the following year, Sir William Herschel announced to the Royal Society, in two memorable papers read before that body, that there exist sidereal systems consisting of two stars revolving about each other in regular orbits, and constituting what he called binary stars, to distinguish them from double stars, generally so called, in which no such periodic change of position is discoverable. Both the individuals of a binary star are at the same distance from the eye in the same sense in which the planet Uranus and its attendant satellites are said to be at the same distance.

More recent observation has fully confirmed these remarkable discoveries. In 1866, Mr. Brothers published a catalogue of 155 stars of this class, and since that time many more have been discovered, every year increasing the number. Most of these stars require the best telescopes for their observation, being generally so close as to require very high magnifying powers.

703. Extension of the law of gravitation to the stars.-The moment the revolution of one star round another was ascertained, the idea of the possible extension of the great principle of gravitation to these remote regions of the universe naturally suggested itself. Newton has proved in his Principia, that if a body revolve in an ellipse by an attractive force directed to the focus, that force will vary according to the law which characterises gravitation. Thus an elliptical orbit became a test of the presence and sway of the law of gravitation. If, then, it could be ascertained that the orbits of the double stars were ellipses, we should at once arrive at the fact that the law of which the discovery conferred such celebrity on the name of Newton, is not confined to the solar system, but prevails throughout the universe.

704. Orbit of star around star elliptic. The first distinct system of calculation by which the true elliptic elements of the orbit of a binary star were ascertained, was supplied in 1830, by M. Savary, who showed that the motion of one of the most remarkable of these stars (Ursæ majoris) indicated an elliptic orbit described in 58 years. Professor Encke, by another process, arrived at the fact that the star 70 Ophiuchi moved in an ellipse with a period of 74 years. Several other orbits were ascertained

and computed by Sir John Herschel, MM. Mädler, Hind, Smyth, and others.

705. Remarkable case of y Virginis.— The most remarkable of these, according to Sir John Herschel, is y Virginis; not only on account of the length of its period, but by reason also of the great diminution of apparent distance and rapid increase of angular motion about each other, of the individuals composing it. It is a bright star of the fourth magnitude, and its component stars are almost exactly equal. It has been known to consist of two stars since the beginning of the eighteenth century, their distance being then between six and seven seconds; so that any tolerably good telescope would resolve it. Since that time they have been in conjunction, so that no telescope that was not of very superior quality was competent to show them otherwise than as a single star somewhat lengthened in one direction. At the present time the stars have passed each other, their angular distance being about four seconds. It fortunately happens that Bradley, in 1718, noticed and recorded, in the margin of one of his observationbooks, the apparent direction of their line of junction as being parallel to that of two remarkable stars a and 8 of the same constellation, as seen by the naked eye. They are entered also as distinct stars in Mayer's catalogue; and this affords also another means of recovering their relative situation at the date of his observations, which were made about the year 1756. Without particularising individual measurements, which will be found in their proper repositories, it will suffice to remark, that their whole series is represented by an ellipse.

706. Singular phenomena produced by one solar system thus revolving round another. To understand the curious effects which must attend the case of a lesser sun with its attendant planets revolving round a greater, let the larger sun, fig. 100, with its planets be represented as s, in the focus of an ellipse, in which the lesser sun accompanied by its planets moves. At A this latter sun is in its perihelion, and nearest to the greater sun s. Moving in its periodical course to B, it is at its mean distance from the sun s. At D it is at aphelion, or its most distant point, and finally returns through c to its perihelion A. The sun s, because of its vast distance from the system A, would appear to the inhabitants of the planets of the system a much smaller than their proper sun; but, on the other hand, this effect of distance would be to a certain extent compensated by its greatly superior magnitude; for analogy justifies the inference that the sun s is greater than the sun A in a proportion equal to that of the magnitude of our sun to one of the planets. The inhabitants of the planets of the system a will then behold the spectacle of two suns in their

firmament. The annual motion of one of these suns will be determined by the motion of the planet itself in its orbit, but that of

Fig. 100.

the other and more distant sun will be determined by the period of the lesser sun around the greater in the orbit A B D C. The rotation of the planets on their axes will produce two days of equal length, but not commencing or ending simultaneously. There will be in general two sunrises and two sunsets! When a planet is situate in the part of its orbit between the two suns, there will be no night. The two suns will then be placed exactly as our sun and moon are placed when the moon is full. When the one sun sets, the other will rise; and when the one rises, the other will set. There will be, therefore, continual day. On the other hand, when a planet is at such a part of its orbit that both suns lie in nearly the same direction as seen from it, both suns will rise and both will set together. There will

then be the ordinary alternation of day and night as on the earth, but the day will have more than the usual splendour, being enlightened by two suns.

In all intermediate seasons the two suns will rise and set at different times. During a part of the day both will be seen at once in the heavens, occupying different places, and reaching the meridian at different times. There will be two noons. In the morning for some time, more or less, according to the season of the year, one sun only will be apparent, and in like manner, in the evening, the sun which first rose will be the first to set, leaving the dominion of the heavens to its splendid companion.

The diurnal and annual phenomena incidental to the planets attending the central sun s will not be materially different, except that to them the two suns will have extremely different magni