minutely subdivided; but their large correspondence with facts remains unaltered. We may briefly remind our readers of their distinctive peculiarities.* The first type consists of white stars, with strongly absorbent hydrogen atmospheres (Sirius, Vega, Rigel, are examples); the second, of solar stars, in which fine metallic rulings are prominent; the third, of red stars like Antares and Betelgeux, showing banded spectra of undetermined chemical origin; the fourth, of redder stars, their photospheres closely veiled by carbon vapours. This mode of grouping, then, remains essentially fixed; and has proved a help, rather than a hindrance, towards the evaluation of minor differences. One of these is especially important. It depends upon the representation of helium in stellar spectra, and establishes an unmistakeable partition within the first type between 'Orion' and Sirian' stars (as they may be conveniently designated). In the former kind, which prevails almost exclusively in the constellation Orion, helium displays conspicuous dark lines; in the latter, patterned by Sirius, they are scarcely, if at all, perceptible. In the solar spectrum, as is well known, helium makes no sign, notwithstanding the profusion with which it occurs in the solar vicinity; so that its disappearance from Sirian spectra does not imply its absence from Sirian atmospheres. Now helium' or Orion stars merge imperceptibly into Sirians, Sirian into solar, solar into Antarian' stars.† There is no gap in the ranks; no line of demarcation can be drawn between them. These four great stellar orders are linked together by small marginal variations into a chain that cannot be broken. Their arrangement points obviously to continuous progression. But the progression might conceivably be in either direction-forward, from white to red suns, or backward, from red to white. We have, then, to seek out some trustworthy criterion by which to distinguish the beginning from the end of the long process of stellar developement. Looking at the matter in its wider bearings, we find two such available. One hundred and nine years ago, the conviction that nebulæ are inchoate stars was forced upon Herschel by the irresistible logic of the eye. He met with them in every stage of condensation; he met with stars provided with See No. 312 of this Journal, p. 415. To borrow Sir Norman Lockyer's convenient designation for third-type stars. Nature, May 18, 1899. every form of nebular appurtenance. His experience, and the conclusion founded upon it, have since been amply confirmed. The lucid matter' of space is neither more nor less than star-spawn. Small compact nebula-those called "planetary' among others--may give rise to single stars; while great nebulæ seem to be nurseries of entire stellar systems. Striking evidence to this effect is adduced by Dr. Roberts in the fine work the title of which is quoted among the headings of this article. His argument is pictorial. The perpetual recurrence, in all quarters of the sky, of curves and lines of equal stars, regularly interspaced, is attested by his plates; he next points to similar configurations, likewise self-delineated, of stars just connected by thin threads of nebulous matter; finally, to corresponding arrangements of nebulous nodules along the structural groovings of vast hazy formations; and draws the plausible inference that such nodules are embryo stars, which grow by consuming, and eventually survive, the nebulæ they are embedded in. And this Herschelian method of demonstration by the continuity of instances gains much stringency through the substitution of photographic for visual comparisons. We may here pause momentarily to congratulate Dr. Roberts on his achievements in celestial photography during the last fifteen years. His leading aim has been the portrayal of nebulæ, and his results, published in two quarto volumes, dated 1894 and 1900 respectively, are a triumph of steady persistence over climatic disadvantages and discouragements. They are noteworthy for what they comprehend no less than for what they exhibit and disclose. They may, indeed, be said to initiate that photographic survey of the nebular heavens which will doubtless be carried out to completion in the twentieth century. What it will amount to can be gathered from the estimate of one hundred and twenty thousand as the number of nebulæ lying within the photographic grasp of the three-foot Crossley reflector, lately mounted at the Lick Observatory in California. This instrument is of the same type as that in use at Crowborough, but greatly exceeds it in light-gathering power. To resume. Since stars develope out of nebulæ, those with patent nebular relationships must be still in the formative era of existence. Their spectra, accordingly, belong to the earliest type. And it proves to be of the 'Orion pattern. Helium-absorption seems invariably to characterise stars just emerging from the nebular matrix. As examples we may take the Pleiades, the entanglement of which in nebulæ of all sorts and conditions is one of the most singular revelations of the camera. Nearly all the Orion brilliants, too, have nebular attachments, more or less close and obvious. The famous nebula in the Sword-handle is only the nucleus of a formation embracing wellnigh the entire asterism. Of prodigious extent and complexity, it constitutes a subordinate universe still partially chaotic. Within its compass, however, some luminous vortices have eddied into suns'-such colossal suns as Rigel, Bellatrix, and the three in the Belt-the light of which, in Professor Pickering's opinion, probably takes a thousand years to reach us. Some of the Orion stars make local centres of concentration, and own individual misty appendages, while the nebulous affinities of others are more general and less visually apparent; but that stars and nebulæ here combine to form a system on a prodigious scale, and of inconceivable organisation, can admit of no manner of doubt. And the stars are of the helium variety. So much for the application of our first criterion. That of the second is less easy, but leads virtually to the same conclusion. The principle involved is thus laid down by Sir William and Lady Huggins : 'Successive evolutional stages of a star's life in the order of their sequence,' they write, must clearly be along the line of changes which a star passes through in consequence of the gradual exhaustion of its original store of energy by radiation. In a classification of stars, that type of star must come first which we have reason to believe to be the most diffuse, or, in other words, in the stage in which condensation is least advanced. Relative density, as representing relative richness in potential energy, should be taken as the guiding principle of a natural system of classification. It seems clear that the earliest stars are not to be regarded as the hottest.' ('Atlas of Spectra,' p. 76.) Nothing in cosmical history is, in truth, more certain than that luminous bodies condense with the efflux of time. For the steady pull of gravity compels the mutual approach of their particles, as the resistent energy imparted to them by heat is slowly squandered by radiation. Two forces are pitted against each other, one unwasting, its opponent unrecuperative. Dissipation of energy' appears, to our limited apprehension, to be the law of the universe, and pari passu with its advance gravity infallibly gains the upper hand. Thus our sun was assuredly more tenuous a million years ago, and will assuredly be more compact a million years hence, than it is now. And the same may be said of every star in the heavens. We cannot, of course, infer an individual star to be younger than another merely because it is specifically lighter, since primitive differences of constitution might largely modify single results. But if we find a low mean density to be a distinguishing feature of a class of stars, we may securely conclude that the objects it comprises are near the outset of their careers radiators. as Now stellar eclipses afford a ready means of fixing, at any rate an upper limit to the density of the bodies undergoing them. Their specific gravity-so to call it cannot exceed, although it may fall short of, a certain assigned value. This value proves to be extremely small. Algol-variables,' taken all round, are of less than one quarter the solar density; and the sun is but slightly heavier than an equally large globe of water, notwithstanding its metallic constitution and the tremendous pressure of its voluminous strata. Transcendent internal temperature is thus indicated. A word about eclipsing stars. They are recognised as such by the sudden failures of light which overtake them at regular intervals of a few days or hours, while outside of these minima' they shine with sensible constancy. Each thus shows itself to be a close binary system, revolving very nearly in our visual plane. And here the spectroscope comes in most opportunely to ratify what had already been less directly asserted by the photometer. For by its means the progress is demonstrated of radial movements corresponding exactly to those assumed on the theory of occultations. Algol, in the head of Medusa, is the model of the class, which now includes nearly a score of members. It has a period of revolution of less than three days. Its eclipses occupy about ten hours, and they are produced by the interposition of a dark globe almost the size of our sun. other cases a combination of two bright globes gives rise to similar, but easily distinguishable effects. The more attentively, in short, these wonderful objects are studied, the more various and interesting are found to be their characteristic phenomena. In They are marked out as in an early stage of growth, not only by their diffuseness, but also-should the views advanced by Dr. See, of Washington, prove sound-by the shortness of their circulatory periods. If double stars originate by the cleavage of a single primitive mass, and separate through the reactive effects of tidal friction, then the most nearly contiguous are the youngest. To this category Algol-variables obviously belong; moreover, all, so far as present knowledge extends, give first-type spectra, and the two most closely scrutinised-those of Algol and A Tauri-are included in the Orion division. And, remarkably enough, just these couples have, according to the recent calculations of Mr. H. M. Russell, a mean density practically the same, and only one-seventh that of water. Air compressed to one-hundredth of its ordinary volume would fairly represent their average consistence. The circumstance is surely significant that just the two helium stars amenable to a determinate test should prove to be of such extreme tenuity. The inference of a primitive condition for this stellar class, arrived at on other grounds, is, then, so far ratified, and may be admitted without hesitation. This suffices to establish a starting-point from which the procession of types advances, with a certain inevitableness, through the ages. Not but that incidental modifications must arise; but they do not affect the general character of the changes supervening through slow cooling and contraction. These, as already stated, lead without a break, by subtle gradations, from Orion to Sirian and solar stars, and onward to the Antarian family. In Sir William Huggins's words: 'The white stars, which are the most numerous, represent the early adult and most persistent stage of stellar life, the solar condition that of maturity and commencing age, while in the orange and the red stars we see the setting-in and advance of old age.' † Here we are brought to a halt. Many stars with banded spectra-Antares, Betelgeux, y Crucis-shine splendidly, despite the vaporous veils that impede and redden their beams. They are vastly, if not immeasurably, remote, and must hence be regarded as suns much more potent than our own. Assuredly no shadow of coming extinction has as yet fallen upon them. Yet it cannot be far off. How, it may be asked, are we to recognise its approach? What are the symptoms of actual stellar decadence? The conjecture may be hazarded that 'long-period variables' supply the missing link with totally dark stars. These extraordinary objects are characterised by banded spectra of both varieties; they hibernate in obscurity usually during *Astrophysical Journal, December 1899, p. 317. |
