recognise three distinct layers, the rain clouds, the cumulus or woolpack clouds, and the cirrus or feathery snow clouds, besides two or three subordinate formations, as the stratus, cumulo-stratus, and cirrostratus; probably many more formations would exist in the cloudladen atmospheres of the giant planets, while assuredly each layer would be very much denser and very much deeper than the corresponding layers in our own atmosphere. Now, the apparent surface of a planet in this condition-that by which its volume would be determined—would be the outer surface of the outermost cloud-layer; if the layers were numerous and deep, this outermost layer would be so far from the real surface that the volume thus determined would be far in excess of the planet's true volume. The mean density inferred from this erroneous determination of the volume would be far less than the planet's true mean density. Now, we find that the mean density of Jupiter is but about one-fourth that of the earth, while that of Saturn is even less, being but about one-seventh of the earth's. We note further, that whereas, if those giant planets were in the same state as the earth, they would most probably be denser than she is, they are less dense, precisely as they would be if still in a state of intense heat. We thus seem forced to the conclusion that they really are in a state of intense heat, apart from that à priori reasoning which had led us to anticipate as much. The agreement between our à priori reasoning and the observed facts adds greatly, it need hardly be said, to the force of the inference, which might be safely enough deduced from either separately.

Let us next consider the direct observational evidence of intense disturbance in the cloud-laden atmosphere. We note that any disturbance on Jupiter, which could be recognised from the earth, must take place on a very large scale. A surface as large as that of England would be quite imperceptible in our best telescopes at Jupiter's distance. The moons of Jupiter appear little more than points in the telescope, and when they are passing over his disc they are scarcely to be discerned at all, unless they happen to be on a portion of his surface having a very different lustre from theirs. Yet, the least of these moons hides a surface of more than three millions of square miles. Probably the smallest perceptible marking on Jupiter would correspond to a portion of Jupiter's surface not less than a hundred thousand square miles in extent. Saturn being about twice as far away (comparing the distances when either planet is most favourably situated for observation), shows all the details of his surface on a correspondingly reduced scale. Not only so, but he is much less brightly illuminated than Jupiter.

Probably a portion

of Saturn's surface distinctly recognisable from the rest, owing to some difference of tint or lustre, must (even when our best telescopes are used) have a surface of not less than half-a-million of square miles. It will be obvious, then, that a disturbance in Jupiter, and still more in Saturn, to be recognisable from the earth, must take place on a scale incomparably greater than that on which any terrestrial disturbances, even the most tremendous earth-throes, have taken place within the knowledge of man. Over a region hundreds of thousands of square miles in extent, the glowing surface of the planet must be torn by subplanetary forces. Vast masses of in

tensely hot vapour must be poured forth from beneath, and, rising to enormous heights, must either sweep away the enwrapping mantle of cloud which had concealed the disturbed surface, or must itself form into a mass of cloud, recognisable because of its enormous extent, and because its texture differs from that of the cloud masses surrounding it. Such a disturbance, extending in the case of Jupiter over an area as large as France, or in the case of Saturn over an area as large as Russia, would be just discernible with our most powerful telescopes. It might very well be, then, that the surface of either planet should appear absolutely at rest, while yet disturbances of the most tremendous character were taking place in every part of the planet's globe. If over a thousand different regions, each as large as Yorkshire, the whole surface were to change from a condition of rest to such activity as corresponds with the tormented surface of seething metal, and vast clouds formed over all such regions so as to hide the actual glow of the surface, our most powerful telescopes would fail to show the faintest trace of change. And Saturn might be still more tremendously disturbed without our seeing any signs of it.

Or again, if we consider the apparent outline of either planet, and inquire what changes would have to take place in the cloud envelopes near the apparent edge of the disc, to be discernible from the earth, we find again that the changes would have to be so tremendous that we might well despair of over discerning the slightest traces of their occurrence. The diameter of Jupiter is, roughly, about 80,000 miles; and anyone who has ever examined the planet with a powerful telescope knows well that a difference of level in any part of the outline by such an amount as the fortieth or the fiftieth part of the diameter (i.e., by 2,000 or 1,600 miles) would not be discernible. Yet, what a disturbance would be implied by such a change of apparent level if the planet had a surface like that of our earth! If we consider that in the most tremendous

earthquake ever known upon earth, a surface of a few thousand square miles rises or falls by a few yards only, we shall be able to form some idea of the fearful nature of a disturbance by which a surface of several millions of square miles would rise or fall through more than a thousand miles (more than half the distance which separates the surface of the earth from the centre). We cannot imagine that any such disturbances take place in Jupiter or Saturn; but even when we take into account the probability that the outline we see is that of cloud masses, we can scarcely expect to find any discernible change in this outline, when we remember on how enormous a scale the cloud envelope must be disturbed (both laterally and vertically) for the telescopist to recognise any perceptible change.

Thus, when we find in the case of both planets very marked changes of both kinds—large spots forming and disappearing on the surface, and the outline of the disc perceptibly changing in positionwe are forced to conclude that the most tremendous forces are at work beneath the cloud envelopes which form the visible surface of these planets.

There could be no more remarkable illustration of the former class of changes than the appearance of the great red spot which is still visible, and has been visible for more than two years, in the southern hemisphere of Jupiter. With a surface equal to threefourths of the entire surface of the earth, this great oval spot has exhibited changes of form and position only explicable on the supposition that the most remarkable changes are taking place in the whole region occupied by the spot. There must be all over that region an activity of disturbance far greater than we have on this earth over the comparatively minute regions disturbed by our fiercest cyclones. The ruddy lustre of this region can hardly be explained, except on the assumption that the light coming from it is partly due to the intense heat, not indeed of the surface here seen, but more probably of regions below that surface. The mere formation of such a spot (remembering always its enormous extent) would imply intense activity in the planet; but its continuing so long visible, while all the time undergoing changes which, though slow in appearance, are in reality stupendous, shows that this part, at any rate, of the planet is the scene of disturbances utterly unlike any which are taking place on our earth.

On Saturn there have been spots and other markings which, though not so remarkable as the great spot on Jupiter, have been quite sufficient, especially when the much greater distance of Saturn

is taken into account, to establish the occurrence of disturbances utterly inconsistent with the idea that Saturn is a habitable globe.

But it is when we examine the changes which have taken place in the outline of the giant planets that we perceive how unlike these orbs are to our own earth.

Take first the distortion of Saturn, which has been called the planet's "square-shouldered aspect." Seen by Sir W. Herschel in 1805 with three different telescopes, and then made the subject of measurement, this peculiarity of figure might, from that series of observations alone, be accepted as indicating a real objective change. But the distortion, together with others as remarkable, has been observed by Sir John Herschel, by the Bonds, in America—than whom no better observers ever lived—and by Coolidge, another American observer. It has been more than once observed by Sir George Airy. And even those unimaginative persons-or rather, those persons whose duty it is to set imagination altogether on one side-the regular observers at Greenwich, record as calmly as one might note that a cloud had changed in form, that "from time to time this year" (1865, I think it was) the planet Saturn has assumed the square-shouldered aspect. We are bound to believe that the planet's outline had really changed. Of course, no solid surface had risen or sunk to the enormous distance, and over the enormous extent of surface, necessary to produce the apparent change. But layers of clouds must for the time have formed above the sub-tropical zones of Saturn, at a height great enough to produce the apparent bulging out of the globe along those zones. Or else the equatorial cloudzones must for the time have changed from the form of visible cloud to that of invisible vapour. Or, more probably, changes of both kinds have taken place. But although such cloud-changes are far less wonderful to think of than changes of equal range in the level of the planet's solid surface, they imply tremendous activity, produced, no doubt, by tremendous heat.

In the case of Jupiter, though Schröter notices occasional flattening of the outline of the disc, which, though slight in appearance, would, if real, have involved great changes in the planet's cloud envelopes, we have no satisfactory evidence of the kind just described in Saturn's case. Schröter, though a careful observer, may have been deceived, and no others have noted such apparent changes of form. But we have evidence of another kind which is, if possible, even more convincing.

The outline of Jupiter is ordinarily estimated by the eye without any extraneous means of measurement. Indeed, even such measure

ment as Sir W. Herschel applied to Saturn is insufficient to detect the slight differences of level which seem to be indicated by observations like Schröter's. But there are occasions when disc-measurers far more trustworthy than any instruments men could devise come into positions enabling us to recognise, at any rate, changes of particular kinds. These are the satellites, which pass alternately in front of Jupiter's disc and behind it, nearly always (in the case of the inner satellites always) transiting the disc and being hidden (or occulted) by it once in each revolution around the planet. Now, theoretically we could recognise changes in the outline of the disc by careful observation of the time at which transit or occultation began or ended. For clearly, if the outline were unusually expanded or contracted where a satellite approached it, the transit would seem to begin earlier or later, respectively; if there were expansion or contraction when the satellite was about to leave the disc, transit would seem to end later or sooner respectively; and there would be corresponding time differences for the beginning and end of occultation. But as a matter of fact the observation of the times of entrance and exit, &c., is too delicate to be available, in the present position of observational astronomy, in this manner. But if it so chanced that a change of level were taking place at the moment when transit or occultation began or ended, such change occurring at the spot where the satellite was, then we might fairly hope that, owing to the proximity (apparent only, of course) of the satellite, the change in the outline of the disc might be detected. It would only be on very rare occasions that this could be expected, but clearly it might happen, if the giant planets are in the condition we have surmised.

Now, just such a case is recorded by Admiral Smyth as having happened in 1828. Here is the account given by a writer by no means too imaginative-Webb, in his "Celestial Objects": "The most surprising" (we would give the noun-substantive, but there is none, probably "observation," or "peculiarity," or something of that sort) "the most surprising is a phenomenon which requires and possesses the highest attestation. 1828, June 26, II." (the second satellite counting from the planet) "having fairly entered on Jupiter, was found twelve or thirteen minutes afterwards outside the limb, where it remained visible for at least four minutes and then suddenly vanished.” The authority of such an observer as Smyth would alone have established this wonderful fact; but it was recorded by two other very competent witnesses, and (what is especially remarkable) at considerable distances, Maclear at twelve miles, and Pearson at thirty-five miles, from Smyth at Bedford. Explanation is here set at defiance;