ginning of 1838 it suddenly increased in lustre so as to surpass all the stars of the first magnitude except Sirius, Canopus, and a Centauri, which last star it nearly equalled. Thence it again diminished, but this time not below the first magnitude until April, 1843, when it had again increased so as to surpass Canopus, and nearly equal Sirius in splendour. "A strange field of speculation," it has been remarked, "is opened by this phænomenon. The temporary stars heretofore recorded have all become totally extinct. Variable stars, so far as they have been carefully attended to, have exhibited periodical alternations, in some degree at least regular, of splendour and comparative obscurity. But here we have a star fitfully variable to an astonishing extent, and whose fluctuations are spread over centuries, apparently in no settled period, and with no regularity of progression. What origin can we ascribe to these sudden flashes and relapses? What conclusions are we to draw as to the comfort or habitability of a system depending for its supply of light and heat on so uncertain a source?" Speculations of this kind can hardly be termed visionary, when we consider that, from what has before been said, we are compelled to admit a community of nature between the fixed stars and our own sun; and when we reflect that geology testifies to the fact of extensive changes having taken place at epochs of the most remote antiquity in the climate and temperature of our globe; changes difficult to reconcile with the operation of secondary causes, such as a different distribution of sea and land, but which would find an easy and natural explanation in a slow variation of the supply of light and heat afforded primarily by the sun itself.

(831.) The Chinese annals of Ma-touan-lin*, in which stand officially recorded, though rudely, remarkable astronomical phænomena, supply a long list of "strange stars," among which, though the greater part are evidently comets, some may be recognized as belonging in all probability to the class of Temporary Stars as above characterized. Such is that which is recorded to have appeared in A. D. 173, between a Translated by M. Edward Biot, Connoissance des Temps, 1846.

and B Centauri, which (no doubt, scintillating from its low situation) exhibited "the five colours," and remained visible from December in that year till July in the next. And another which these annals assign to A. D. 1011, and which would seem to be identical with a star elsewhere referred to A. D. 1012, "which was of extraordinary brilliancy, and remained visible in the southern part of the heavens during three months," a situation agreeing with the Chinese record, which places it low in Sagittarius. Among several less unequivocal is one referred to B. C. 134, in Scorpio, which may possibly have been Hipparchus's star. None of the stars of A. D. 389, 945, 1264, and 1572, however, are noticed in these records. It is worthy of especial notice, that all the stars of this kind on record, of which the places are distinctly indicated, have occurred, without exception, in or close upon the borders of the Milky Way, and that only within the following semicircle, the preceding having offered no example of the kind.

(832.) On a careful re-examination of the heavens, and a comparison of catalogues, many stars are now found to be missing; and although there is no doubt that these losses have arisen in the great majority of instances from mistaken entries, and in some from planets having been mistaken for stars, yet in some it is equally certain that there is no mistake in the observation or entry, and that the star has really been observed, and as really has disappeared from the heavens. The whole subject of variable stars is a branch of practical astronomy which has been too little followed up, and it is precisely that in which amateurs of the science, and especially voyagers at sea, provided with only good eyes, or moderate instruments, might employ their time to excellent advantage. It holds out a sure promise of rich discovery, and is one in which astronomers in established observatories are almost of necessity precluded from taking a part by the nature of the observations required. Catalogues of the com

* Hind, Notices of the Astronomical Society, viii. 156., citing Hepidannus. He places the Chinese star of 173 B.C. between a and B Canis Minoris, but M. Biot distinctly says a, ß pied oriental du Centaure.

parative brightness of the stars in each constellation have been constructed by Sir Wm. Herschel, with the express object of facilitating these researches, and the reader will find them, and a full account of his method of comparison, in the Phil. Trans. 1796, and subsequent years.

(833.) We come now to a class of phænomena of quite a different character, and which give us a real and positive insight into the nature of at least some among the stars, and enable us unhesitatingly to declare them subject to the same dynamical laws, and obedient to the same power of gravitation, which governs our own system. Many of the stars, when examined with telescopes, are found to be double, i. e. to consist of two (in some cases three or more) individuals placed near together. This might be attributed to accidental proximity, did it occur only in a few instances; but the frequency of this companionship, the extreme closeness, and, in many cases, the near equality of the stars so conjoined, would alone lead to a strong suspicion of a more near and intimate relation than mere casual juxtaposition. The bright star Castor, for example, when much magnified, is found to consist of two stars of nearly the third magnitude, within 5′′ of each other. Stars of this magnitude, however, are not so common in the heavens as to render it otherwise than excessively improbable that, if scattered at random, they would fall so near. But this improbability becomes immensely increased by a consideration of the fact, that this is only one out of a great many similar instances. Mitchell, in 1767, applying the rules for the calculation of probabilities to the case of the six brightest stars in the group called the Pleiades, found the odds to be 500000 to 1 against their proximity being the mere result of a random scattering of 1500 stars (which he supposed to be the total number of stars of that magnitude in the celestial sphere) over the heavens. Speculating further on this, as an indication of physical connexion rather than fortuitous assemblage, he was led to surmise the possibility (since converted into a certainty, but at that time,

This number is considerably too small, and in consequence, Mitchell's odds in this case materially overrated. But enough will remain, if this be rectified, fully to bear out his argument. Phil. Trans. vol. 57.

antecedent to any observation) of the existence of compound stars revolving about one another, or rather about their common center of gravity. M. Struve, pursuing the same train of thought as applied specially to the cases of double and triple combinations of stars, and grounding his computations on a more perfect enumeration of the stars visible down to the 7th magnitude, in the part of the heavens visible at Dorpat, calculates that the odds are 9570 to 1 against any two stars, from the 1st to the 7th magnitude inclusive, out of the whole possible number of binary combinations then visible, falling (if fortuitously scattered) within 4" of each other. Now the number of instances of such binary combinations actually observed at the date of this calculation was already 91, and many more have since been added to the list. Again, he calculates that the odds against any such stars fortuitously scattered, falling within 32" of a third, so as to constitute a triple star, is not less than 173524 to 1. Now, four such combinations occur in the heavens; viz. Orionis, a Orionis, 11 Monocerotis, and Cancri. The conclusion of a physical connexion of some kind or other is therefore unavoidable.

(834.) Presumptive evidence of another kind is furnished by the following consideration. Both a Centauri and 61 Cygni are "Double Stars." Both consist of two individuals, nearly equal, and separated from each other by an interval of about a quarter of a minute. In the case of 61 Cygni, the stars exceeding the 7th magnitude, there is already a primâ facie probability of 9578 to 1 against their apparent proximity. The two stars of a Centauri are both at least of the 2nd magnitude, of which altogether not more than about 50 or 60 exist in the whole heavens. But, waving this consideration, both these stars, as we have already seen, have a proper motion so considerable that, supposing the constituent individuals unconnected, one would speedily leave the other behind. Yet at the earliest dates at which they were respectively observed these stars were not perceived to be double, and it is only to the employment of telescopes magnifying at least 8 or 10 times, that we owe the knowledge we now possess of their being so. With such a telescope Lacaille, in

1751, was barely able to perceive the separation of the two constituents of a Centauri, whereas, had one of them only been affected with the observed proper motion, they should then have been 6′ asunder. In these cases then some physical connexion may be regarded as proved by this fact alone.

(835.) Sir William Herschel has enumerated upwards of 500 double stars, of which the individuals are less than 32" asunder. M. Struve, prosecuting the inquiry with instruments more conveniently mounted for the purpose, and wrought to an astonishing pitch of optical perfection, has added more than five times that number. And other observers have extended still further the catalogue of "Double Stars," without exhausting the fertility of the heavens. Among these are a great many in which the distance between the component individuals does not exceed a single second. They are divided into classes by M. Struve (the first living authority in this department of Astronomy) according to the proximity of their component individuals. The first class comprises those only in which the distance does not exceed 1"; the 2nd those in which it exceeds 1" but falls short of 2"; the 3rd class extends from 2" to 4" distance; the 4th from 4" to 8"; the 5th from 8" to 12"; the 6th from 12" to 16"; the 7th from 16" to 24", and the 8th from 24" to 32". Each class he again subdivides into two sub-classes of which the one under the appellation of conspicuous double stars (duplices lucida) comprehends those in which both individuals exceed the 8 magnitude, that is to say, are separately bright enough to be easily seen in any moderately good telescope. All others, in which one or both the constituents are below this limit of easy visibility, are collected into another sub-class, which he terms residuary (Duplices reliquæ). This arrangement is so far convenient, that after a little practice in the use of telescopes as applied to such objects, it is easy to judge what optical power will probably suffice to resolve a star of any proposed class and either sub-class, or would at least be so if the second or residuary sub-class were further subdivided by placing in a third sub-class "delicate" double stars, or those in which the companion star is so very minute as to