observation depends need not here be explained, because they have already been considered at some length in a paper entitled the "Gamut of Light," in my treatise, "Orbs around us." But I may so far recapitulate what I have there said, as to note that if we are approaching Sirius or receding from him, either through his motion or the sun's, or through the combined effects of both motions, the waves of light which travel to us from Sirius must appear shortened or lengthened, precisely as seawaves would seem narrower or broader according as a swimmer travelled against or with their onward course. Now the light from a star contains all degrees of wavelength from the longest light-waves (which correspond to the red end of the spectrum) down to the shortest (which correspond to the violet end); so that amidst all these wave-lengths the observer could no more recognise such a change as would result from approach or recess than the swimmer of our illustrative case could recognise the apparent shortening or lengthening of waves in a stormtost sea where waves of all dimensions were abroad. But if light-waves of any specified length can be in any way distinguished from the rest, the case (as respects them) corresponds to that of a swimmer crossing a long and uniform succession of rollers. Now the dark lines in the spectrum of a star, when they can be certainly identified with the lines belonging to the spectrum of some known element, supply this very knowledge of the true wavelengths. Dr Huggins had identified certain very well marked lines in the spectrum of Sirius with the well

M

known lines of hydrogen. If he could find that these lines in the star's spectrum are measurably displaced either towards the red or the violet end of the spectrum, he could infer that the wave-lengths of the star's light are measurably lengthened or shortened through a recession or approach on the part of the star. This he actually effected. He found that one of the hydrogen lines of the star was displaced in such a way as to indicate a lengthening of the light-waves corresponding to a recession at the rate of forty-one miles per second. But a part of this recession was due to the earth's orbital motion at the time of observation, and another part is due to the sun's own motion through space. There remains, after these portions have been deducted, a motion of recession in space amounting to about twenty-six miles per second. This rate of motion-or rather a recession from the sun at the rate of twenty-nine miles per second-is absolute, not being affected in any way by our estimate of the distance of Sirius. Combining the recession in space with the estimated thwart motion of twenty miles per second, we deduce a real motion in space amounting to about thirtythree miles per second.*

But the circumstance which remains to be mentioned respecting Sirius before this paper is drawn to a conclusion, is perhaps more remarkable than any yet referred to.

* Of course the two motions must not be simply added together, since they are not in the same direction. The actual motion is represented by the diagonal of an oblong whose sides represent the motion of recession and the thwart motion.

When astronomers compared together the places of Sirius as recorded in a long series of observations, they found what appeared like a periodic displacement of the star. In the first instance, they had examined only the recorded positions of the star as respects east and west; and the observed displacement in this direction suggested that in reality Sirius is circling around another orb, or rather that Sirius and some other orb are circling around a common centre, in a period of fifty years. When it was found that the star appears to drift to north and south of its mean place in a manner according very closely with this hypothesis, astronomers naturally began to regard the theory as rendered highly probable by a coincidence which could scarcely be regarded as accidental.

But no star had been seen where this theory required that a star should be; and moreover the theory required an orb whose bulk should be about two-thirds of the enormous bulk of Sirius, and it was to be inferred that so large an orb would shine with a lustre comparable with that of Sirius himself. On this last point, however, it was well remarked that we have no sufficient reason for believing that all the orbs which people space are luminous. However, a search was instituted for the star which the theory seemed to require. Nor was the search unsuccessful. With a telescope 18 inches in aperture, made by himself, the eminent American optician, Alvan Clark, detected a faint star close by Sirius, apparently, though actually (on a moderate computation), at least 2,000 millions of miles from him. The movements of this star have been

held by some astronomers to accord fairly with the requirements of the theory just mentioned; though I must admit that I fail to find a very close resemblance between the actual motion of the faint companion and those which the theory requires. But we now have a choice of disturbing companions, since the late Mr Goldsmidt (who far surpassed even our own "eagle-eyed Dawes" in keenness of vision) not only saw Clark's star with a telescope only four inches in aperture, but actually succeeded in detecting five other companion stars.

We

We can best explain the faintness of these stars by supposing that they are opaque bodies which shine only by reflecting the light which they receive from their sun Sirius. But if so, they must be globes of enormous real dimensions, the least of them probably exceeding our own sun many times in volume, while the greatest (so we may conclude from the disturbance Sirius himself undergoes) must be so large and massive that a thousand such orbs as our sun would not equal it either in bulk or mass. have here, then, a system differing altogether in character from our solar system, the largest member of which is but equal in mass to about the 1,300th part of the sun. The complete Sirian system, may even outweigh Sirius himself, and its mass added to his must exert an attractive influence throughout an enormous portion of the stellar system. It would seem, indeed, not wholly impossible that Sirius holds a higher rank in the scale of creation that our sun and other similar orbs-that compared with him these are as secondary orbs compared with primaries.

Without insisting on this, however, we may assert with confidence that whether we consider his volume, his bulk, or the mighty energy evidenced by his brightness, Sirius well merits the title under which he has been here described. Of all the orbs with which astronomers have to deal, he seems worthiest to be called par excellence the giant sun.