There has thus been opened up an entirely new field of research, and the idea, long cherished, that the stars might prove to have dark, or, at all events, invisible, companions attendant on them, somewhat as our own sun has its planets, has been proved to be perfectly sound. So far, in the case of dark companions, only bodies of such vast size have been detected as to render any comparison with the planets of our system difficult; but the principle is established, and the probability of great numbers of the stars having real planetary systems attendant on them is so great as to become practically a certainty. 'We naturally infer,' says Professor Newcomb, 'that ... innumerable stars may have satellites, planets, or companion stars so close or so faint as to elude our powers of observation.'
From the consideration of spectroscopic binaries we naturally turn to that of variable stars, the two classes being, to some extent at least, coincident, as is evidenced by the case of Algol. While the discovery of spectroscopic binaries is one of the latest results of research, that of variability among stars dates from comparatively far back in the history of astronomy. As early as the year 1596 David Fabricius noted the star now known as Omicron Ceti, or Mira, 'the Wonderful,' as being of the third magnitude, while in the following year he found that it had vanished. A succession of appearances and disappearances was witnessed in the middle of the next century by Holwarda, and from that time the star has been kept under careful observation, and its variations have been determined with some exactness, though there are anomalies as yet unexplained. 'Once in eleven months,' writes Miss Clerke, 'the star mounts up in about 125 days from below the ninth to near the third, or even to the second magnitude; then, after a pause of two or three weeks, drops again to its former low level in once and a half times, on an average, the duration of its rise.' This most extraordinary fluctuation means that at a bright maximum Mira emits 1,500 times as much light as at a low minimum. The star thus subject to such remarkable outbursts is, like most variables, of a reddish colour, and at maximum its spectrum shows the presence of glowing hydrogen. Its average period is about 331 days; but this period is subject to various irregularities, and the maximum has sometimes been as much as two months away from the predicted time. Mira Ceti may be taken as the type of the numerous class of stars known as 'long-period variables.'
Not less interesting are those stars whose variations cover only short periods, extending from less than thirty days down to a few hours. Of these, perhaps the most easily observed, as it is also one of the most remarkable, is Beta Lyræ. This star is one of the two bright stars of nearly equal magnitude which form an obtuse-angled triangle with the brilliant first-magnitude star Vega. The other star of the pair is Gamma Lyræ, and between them lies the famous Ring Nebula, to be referred to later. Ordinarily Beta Lyræ is of magnitude 3·4, but from this it passes, in a period of rather less than thirteen days, through two minima, in one of which it descends to magnitude 3·9 and in the other to 4·5. This fluctuation seems trifling. It really means, however, that at maximum the star is two and three-quarter times brighter than when it sinks to magnitude 4·5; and the variation can be easily recognised by the naked eye, owing to the fact of the nearness of so convenient a comparison star as Gamma Lyræ. Beta Lyræ is a member of the class of spectroscopic binaries, and belongs to that type of the class in which the mutually eclipsing bodies are both bright. In such cases the variation in brilliancy is caused by the fact that when the two bodies are, so to speak, side by side, light is received from both of them, and a maximum is observed; while, when they are end on, both in line with ourselves, one cuts off more or less of the other's light from us, thus causing a minimum.
A third class, distinct from either of the preceding, is that of the Algol Variables, so-called from the bright star Beta Persei, which has already been mentioned as a spectroscopic binary. Than this star there is no more notable variable in the heavens, and its situation fortunately renders it peculiarly easy of observation to northern students. Algol shines for about fifty-nine hours as a star of small second magnitude, then suddenly begins to lose light, and in four and a half hours has fallen to magnitude three and a half, losing in so short a space two-thirds of its normal brilliancy. It remains in this degraded condition for only fifteen minutes, and then begins to recover, reaching its normal lustre in about five hours more. These remarkable changes, due, as before mentioned, to the presence of an invisible eclipsing companion, are gone through with the utmost regularity, so much so that, as Gore says, the minima of Algol 'can be predicted with as much certainty as an eclipse of the sun.' The features of the type-star are more or less closely reproduced in the other Algol Variables—a comparatively long period of steady light emission, followed by a rapid fall to one or more minima, and a rapid recovery of light. The class as yet is a small one, but new members are gradually being added to it, the majority of them white, like the type-star.
The study of variable stars is one which should seem to be specially reserved for the amateur observer. In general, it requires but little instrumental equipment. Many variables can be seen at maximum, some even at minimum, with the unaided eye; in other cases a good opera or field glass is all that is required, and a 2½ or 3-inch telescope will enable the observer to command quite an extensive field of work. Here, again, the beginner may be referred to the Memoirs of the British Astronomical Association for help and guidance, and may be advised to connect himself with the Variable Star Section.
With the exception of such variations in the lustre of certain stars as have been described, the aspect of the heavens is, in general, fixed and unchanging. There are, as we shall see, real changes of the vastest importance continually going on; but the distances separating us from the fixed stars are so enormous that these changes shrink into nothingness, and the astronomers of forty centuries before our era would find comparatively little change today in the aspect of the constellations with which they were familiar. But occasionally a very remarkable change does take place, in the apparition of a new or temporary star. The accounts of the appearance of such objects are not very numerous, but are of great interest. We pass over those recorded, in more or less casual fashion, by the ancients, for the reason that the descriptions given are in general more picturesque than illuminative. It does not add much to one's knowledge, though it may excite wonder, to find the Chinese annals recording the appearance, in A.D. 173, of a new star 'resembling a large bamboo mat!'
The first Nova, of which we have a really scientific record, was the star which suddenly blazed out, in November, 1572, in the familiar W of Cassiopeia. It was carefully observed by the great astronomer, Tycho Brahé, and, according to him, was brighter than Sirius, Alpha Lyræ, or Jupiter. Tycho followed it till March, 1574, by which time it had sunk to the limit of unaided vision, and further observation became impossible. There is at present a star of the eleventh magnitude close to the place fixed for the Nova from Tycho's observations. In 1604 and 1670, new stars were observed, the first by Kepler and his assistants, the second by the monk Anthelme; but from 1670 there was a long break in the list of discoveries, which was ended by Hind's observation of a new star in Ophiuchus (April, 1848). This was never a very conspicuous object, rising only to somewhat less than fourth magnitude, and soon fading to tenth or eleventh. We can only mention the 'Blaze Star' of Corona Borealis, discovered by Birmingham in 1866, the Nova discovered in 1876 by Schmidt of Athens, near Rho Cygni—an object which seems to have faded out into a planetary nebula, a fate apparently characteristic of this class of star—and the star which appeared in 1885, close to the nucleus of the Great Nebula in Andromeda.
In 1892, Dr. Anderson of Edinburgh discovered in the constellation Auriga a star which he estimated as of fifth magnitude. The discovery was made on January 31, and the new star was found to have been photographed at Harvard on plates taken from December 16, 1891, to January 31, 1892. Apparently this Nova differed from other temporary stars in the fact that it attained its full brightness only gradually. By February 3 it rose to magnitude 3·5, then faded by April 1 to fifteenth, but in August brightened up again to about ninth magnitude. It is now visible as a small star. The great development of spectroscopic resources brought this object, otherwise not a very conspicuous one, under the closest scrutiny. Its spectrum showed many bright lines, which were accompanied by dark ones on the side next the blue. The idea was thus suggested that the outburst of brilliancy was due to a collision between two bodies, one of which, that causing the dark lines, was approaching the earth, while the other was receding from it. Lockyer considered the conflagration to be due to a collision between two swarms of meteorites, Huggins that it was caused by the near approach to one another of two gaseous bodies, while others suggested that the rush of a star or of a swarm of meteorites through a nebula would explain the facts observed. Subsequent observations of the spectrum of Nova Aurigæ have revealed the fact that it has obeyed the destiny which seems to wait on temporary stars, having become a planetary nebula.
Dr. Anderson followed up his first achievement by the discovery of a brilliant Nova in the constellation Perseus. The discovery was made on the night of February 21-22, 1901, the star being then of magnitude 2·7. Within two days it became about the third brightest star in the sky, being a little more brilliant than Capella; but before the middle of April it had sunk to fifth magnitude. The rapidity of its rise must have been phenomenal! A plate exposed at Harvard on February 19, and showing stars to the eleventh magnitude, bore no trace of the Nova. 'It must therefore,' says Newcomb, 'have risen from some magnitude below the eleventh to the first within about three days. This difference corresponds to an increase of the light ten thousandfold!' Such a statement leaves the mind simply appalled before the spectacle of a cataclysm so infinitely transcending the very wildest dreams of fancy. Subsequent observations have shown the usual tendency towards development into a nebula, and in August, 1901, photographs were actually obtained of a nebulosity round the star, showing remarkable condensations. These photographs, taken at Yerkes Observatory, when compared with others taken at Mount Hamilton in November, revealed the startling fact that the condensations of the nebula were apparently in extraordinarily rapid motion. Now the Nova shows no appreciable parallax, or in other words is so distant that its distance cannot be measured; on what scale, therefore, must these motions have been to be recorded plainly across a gulf measurable perhaps in hundreds of light years!
Nova Geminorum, discovered by Professor Turner, at Oxford, in March, 1903, had not the striking features which lent so much interest to Nova Persei. It showed a crimson colour, and its spectrum indicated the presence in its blaze of hydrogen and helium; but it faded so rapidly as to show that the disturbance affected a comparatively small body, and it has exhibited the familiar new star change into a nebula.