PLANETS AND SATELLITES—(continued)

"The analogy between Mars and the earth is perhaps by far the greatest in the whole solar system." So Herschel wrote in 1783,[965] and so we may safely say to-day, after six score further years of scrutiny. The circumstance lends a particular interest to inquiries into the physical habitudes of our exterior planetary neighbour.

Fontana first caught glimpses, at Naples in 1636 and 1638,[966] of dusky stains on the ruddy disc of Mars. They were next seen by Hooke and Cassini in 1666, and this time with sufficient distinctness to serve as indexes to the planet's rotation, determined by the latter as taking place in a period of twenty-four hours forty minutes.[967] Increased confidence was given to this result through Maraldi's precise verification of it in 1719.[968] Among the spots observed by him, he distinguished two as stable in position, though variable in size. They were of a peculiar character, showing as bright patches round the poles, and had already been noticed during sixty years back. A current conjecture of their snowy nature obtained validity when Herschel connected their fluctuations in extent with the progress of the Martian seasons. The inference of frozen precipitations could scarcely be resisted when once it was clearly perceived that the shining polar zones did actually by turns diminish and grow with the alternations of summer and winter in the corresponding hemisphere.

This, it may be said, was the opening of our acquaintance with the state of things prevailing on the surface of Mars. It was accompanied by a steady assertion, on Herschel's part, of permanence in the dark markings, notwithstanding partial obscurations by clouds and vapours floating in a "considerable but moderate atmosphere." Hence the presumed inhabitants of the planet were inferred to "probably enjoy a situation in many respects similar to ours."[969]

SchrÖter, on the other hand, went altogether wide of the truth as regards Mars. He held that the surface visible to us is a mere shell of drifting cloud, deriving a certain amount of apparent stability from the influence on evaporation and condensation of subjacent but unseen areographical features;[970] and his opinion prevailed with his contemporaries. It was, however, rejected by Kunowsky in 1822, and finally overthrown by Beer and MÄdler's careful studies during five consecutive oppositions, 1830-39. They identified at each the same dark spots, frequently blurred with mists, especially when the local winter prevailed, but fundamentally unchanged.[971] In 1862 Lockyer established a "marvellous agreement" with Beer and MÄdler's results of 1830, leaving no doubt as to the complete fixity of the main features, amid "daily, nay, hourly," variations of detail through transits of clouds.[972] On seventeen nights of the same opposition, F. Kaiser of Leyden obtained drawings in which nearly all the markings noted in 1830 at Berlin reappeared, besides spots frequently seen respectively by Arago in 1813, by Herschel in 1783, and one sketched by Huygens in 1672 with a writing-pen in his diary.[973] From these data the Leyden observer arrived at a period of rotation of 24h. 37m. 22·62s., being just one second shorter than that deduced, exclusively from their own observations, by Beer and MÄdler. The exactness of this result was practically confirmed by the inquiries of Professor Bakhuyzen of Leyden.[974] Using for a middle term of comparison the disinterred observations of SchrÖter, with those of Huygens at one, and of Schiaparelli at the other end of an interval of 220 years, he was enabled to show, with something like certainty, that the time of rotation (24h. 37m. 22·735s.) ascribed to Mars by Mr. Proctor[975] in reliance on a drawing executed by Hooke in 1666, was too long by nearly one-tenth of a second. The minuteness of the correction indicates the nicety of care employed. Nor employed vainly; for, owing to the comparative antiquity of the records available in this case, an almost infinitesimal error becomes so multiplied by frequent repetition as to produce palpable discrepancies in the positions of the markings at distant dates. Hence Bakhuyzen's period of 24h. 37m. 22·66s. is undoubtedly of a precision unapproached as regards any other heavenly body save the earth itself.

Two facts bearing on the state of things at the surface of Mars were, then, fully acquired to science in or before the year 1862. The first was that of the seasonal fluctuations of the polar spots; the second, that of the general permanence of certain dark gray or greenish patches, perceived with the telescope as standing out from the deep yellow ground of the disc. That these varieties of tint correspond to the real diversities of a terraqueous globe, the "ripe cornfield"[976] sections representing land, the dusky spots and streaks, oceans and straits, has long been the prevalent opinion. Sir J. Herschel in 1830 led the way in ascribing the redness of the planet's light to an inherent peculiarity of soil.[977] Previously it had been assimilated to our sunset glows rather than to our red sandstone formations—set down, that is, to an atmospheric stoppage of blue rays. But the extensive Martian atmosphere, implicitly believed in on the strength of some erroneous observations by Cassini and RÖmer in the seventeenth century, vanished before the sharp occultation of a small star in Leo, witnessed by Sir James South in 1822;[978] and Dawes's observation in 1865,[979] that the ruddy tinge is deepest near the central parts of the disc, certified its non-atmospheric origin. The absolute whiteness of the polar snow-caps was alleged in support of the same inference by Sir William Huggins in 1867.[980]

All recent operations tend to show that the atmosphere of Mars is much thinner than our own. This was to have been expected À priori, since the same proportionate mass of air would on his smaller globe form a relatively sparse covering.[981] Besides, gravity there possesses less than four-tenths its force here, so that this sparser covering would weigh less, and be less condensed, than if it enveloped the earth. Atmospheric pressure would accordingly be of about two and a quarter, instead of fifteen terrestrial pounds per square inch. This corresponds with what the telescope shows us. It is extremely doubtful whether any features of the earth's actual surface could be distinguished by a planetary spectator, however well provided with optical assistance. Professor Langley's inquiries[982] led him to conclude that fully twice as much light is absorbed by our air as had previously been supposed—say 40 per cent. of vertical rays in a clear sky. Of the sixty reaching the earth, less than a quarter would be reflected even from white sandstone; and this quarter would again pay heavy toll in escaping back to space. Thus not more than perhaps ten or twelve out of the original hundred sent by the sun would, under the most favourable circumstances, and from the very centre of the earth's disc, reach the eye of a Martian or lunar observer. The light by which he views our world is, there is little doubt, light reflected from the various strata of our atmosphere, cloud or mist-laden or serene, as the case may be, with an occasional snow-mountain figuring as a permanent white spot.

This consideration at once shows us how much more tenuous the Martian air must be, since it admits of topographical delineations of the Martian globe. The clouds, too, that form in it seem in general to be rather of the nature of ground-mists than of heavy cumulus.[983] Occasionally, indeed, durable and extensive strata become visible. During the latter half of October, 1894, for instance, a region as large as Europe remained apparently cloud-covered. Yet most recent observers are unable to detect the traces of aqueous absorption in the Martian spectrum noted by Huggins in 1867[984] and by Vogel in 1873.[985] Campbell vainly looked for them,[986] visually in 1894, spectrographically in 1896; Keeler was equally unsuccessful;[987] Jewell[988] holds that they could, with present appliances, only be perceived if the atmosphere of Mars were much richer in water-vapour than that of the earth. There can be little doubt, however, that its supply is about the minimum adequate to the needs of a living, and perhaps a life-nuturing planet.

The climate of Mars seems to be unexpectedly mild. Its theoretical mean temperature, taking into account both distance from the sun and albedo, is 34° C. below freezing.[989] Yet its polar snows are both less extensive and less permanent than those on the earth. The southern white hood, noticed by Schiaparelli in 1877 to have survived the summer only as a small lateral patch, melted completely in 1894. Moreover, Mr. W. H. Pickering observed with astonishment the disappearance, in the course of thirty-three days of June and July, 1892, of 1,600,000 square miles of southern snow.[990] Curiously enough, the initial stage of shrinkage in the white calotte was marked by its division into two unequal parts, as if in obedience to the mysterious principle of duplication governing so many Martian phenomena.[991] Changes of the hues associated respectively with land and water accompanied in lower latitudes, and were thought to be occasioned by floods ensuing upon this rapid antarctic thaw. It is true that scarcity of moisture would account for the scantiness and transitoriness of snowy deposits easily liquefied because thinly spread. But we might expect to see the whole wintry hemisphere, at any rate, frost-bound, since the sun radiates less than half as much heat on Mars as on the earth. Water seems, nevertheless, to remain, as a rule, uncongealed everywhere outside the polar regions. We are at a loss to imagine by what beneficent arrangement the rigorous conditions naturally to be looked for can be modified into a climate which might be found tolerable by creatures constituted like ourselves.

Martian topography may be said to form nowadays a separate sub-department of descriptive astronomy. The amount of detail become legible by close scrutiny on a little disc which, once in fifteen years, attains a maximum of about 1/5000 the area of the full moon, must excite surprise and might provoke incredulity. Spurious discoveries, however, have little chance of holding their own where there are so many competitors quite as ready to dispute as to confirm.

The first really good map of Mars was constructed in 1869 by Proctor from drawings by Dawes. Kaiser of Leyden followed in 1872 with a representation founded upon data of his own providing in 1862-64; and Terby, in his valuable ArÉographie, presented to the Brussels Academy in 1873[992] a careful discussion of all important observations from the time of Fontana downwards, thus virtually adding to knowledge by summarising and digesting it. The memorable opposition of September 5, 1877, marked a fresh epoch in the study of Mars. While executing a trigonometrical survey (the first attempted) of the disc, then of the unusual size of 25' across, G. V. Schiaparelli, director of the Milan Observatory, detected a novel and curious feature. What had been taken for Martian continents were found to be, in point of fact, agglomerations of islands, separated from each other by a network of so-called "canals" (more properly channels).[993] These are obviously extensions of the "seas," originating and terminating in them, and sharing their gray-green hue, but running sometimes to a length of three or four thousand miles in a straight line, and preserving throughout a nearly uniform breadth of about sixty miles. Further inquiries have fully substantiated the discovery made at the Brera Observatory. The "canals" of Mars are an actually existent and permanent phenomenon. An examination of the drawings in his possession showed M. Terby that they had been seen, though not distinctively recognised, by Dawes, Secchi, and Holden; several were independently traced out by Burton at the opposition of 1879; all were recovered by Schiaparelli himself in 1879 and 1881-82; and their indefinite multiplication resulted from Lovell's observations in 1894 and 1896.

When the planet culminated at midnight, and was therefore in opposition, December 26, 1881, its distance was greater, and its apparent diameter less than in 1877, in the proportion of sixteen to twenty-five. Its atmosphere was, however, more transparent, and ours of less impediment to northern observers, the object of scrutiny standing considerably higher in northern skies. Never before, at any rate, had the true aspect of Mars come out so clearly as at Milan, with the 8-3/4-inch Merz refractor of the observatory, between December, 1881, and February, 1882. The canals were all again there, but this time they were—in as many as twenty cases—seen in duplicate. That is to say, a twin-canal ran parallel to the original one at an interval of 200 to 400 miles.[994]

We are here brought face to face with an apparently insoluble enigma. Schiaparelli regards the "germination" of his canals as a periodical phenomenon depending on the Martian seasons. It is, assuredly, not an illusory one, since it was plainly apparent, during the opposition of 1886, to MM. Perrotin and Thollon at Nice,[995] and to the former, using the new 30-inch refractor of that observatory, in 1888; Mr. A. Stanley Williams, with the help of only a 6-1/2-inch reflector, distinctly perceived in 1890 seven of the duplicate objects noted at Milan,[996] and the Lick observations, both of 1890 and of 1892, together with the drawings made at Flagstaff and Mexico during the last favourable oppositions of the nineteenth century, brought unequivocal confirmation to the accuracy of Schiaparelli's impressions.[997] Various conjectures have been hazarded in explanation of this bizarre appearance. The difficulty of conceiving a physical reality corresponding to it has suggested recourse to an optical rationale. Proctor regarded it as an effect of diffraction;[998] Stanislas Meunier, of oblique reflection from overlying mist-banks;[999] Flammarion considers it possible that companion-canals might, under special circumstances, be evoked by refraction as a kind of mirage.[1000] But none of these speculations are really admissible, when all the facts are taken into account. The view that the canals of Mars are vast rifts due to the cooling of the globe, is recommended by the circumstance that they tend to follow great circles; nevertheless, it would break down if, as Schiaparelli holds, the fluctuations in their visibility depend upon actual obliterations and re-emergencies. Fantastic though the theory of their artificial origin appear, it is held by serious astronomers. Its vogue is largely due to Mr. Lowell's ingenious advocacy. He considers the Martian globe to be everywhere intersected by an elaborate system of irrigation-works, rendered necessary by a perennial water-famine, relieved periodically by the melting of the polar snows. Nor does he admit the existence of oceans, or lakes. What have been taken for such are really tracts covered with vegetation, the bright areas intermixed with them representing sandy deserts. And it is noteworthy in this connection that Professor Barnard obtained in 1894,[1001] with the great Lick refractor, "suggestive and impressive views" disclosing details of light and shade on the gray-green patches so intricate and minute as almost to preclude the supposition of their aqueous nature.

The closeness of the terrestrial analogy has thus of late been much impaired. Even if the surface of Mars be composed of land and water, their distribution must be of a completely original type. The interlacing everywhere of continents with arms of the sea (if that be the correct interpretation of the visual effects) implies that their levels scarcely differ;[1002] and Schiaparelli carries most observers with him in holding that their outlines are not absolutely constant, encroachments of dusky upon bright tints suggesting extensive inundations.[1003] The late N. E. Green's observations at Madeira in 1877 indicated, on the other hand, a rugged south polar region. The contour of the snow-cap not only appeared indented, as if by valleys and promontories, but brilliant points were discerned outside the white area, attributed to isolated snow-peaks.[1004] Still more elevated, if similarly explained, must be the "ice island" first seen in a comparatively low latitude by Dawes in January, 1865.

On August 4, 1892, Mars stood opposite to the sun at a distance of only 34,865,000 miles from the earth. In point of vicinity, then, its situation was scarcely less favourable than in 1877. The low altitude of the planet, however, practically neutralised this advantage for northern observers, and public expectation, which had been raised to the highest pitch by the announcements of sensation-mongers, was somewhat disappointed at the "meagreness" of the news authentically received from Mars. Valuable series of observations were, nevertheless, made at Lick and Arequipa; and they unite in testifying to the genuine prevalence of surface-variability, especially in certain regions of intermediate tint, and perhaps of the "crude consistence" of "boggy Syrtes, neither sea, nor good dry land." Professor Holden insisted on the "enormous difficulties in the way of completely explaining the recorded phenomena by terrestrial analogies";[1005] Mr. W. H. Pickering spoke of "conspicuous and startling changes." They, however, merely overlaid, and partially disguised, a general stability. Among the novelties detected by Mr. Pickering were a number of "lakes," or "oases" (in Lowell's phraseology), under the aspect of black dots at the junctions of two or more canals;[1006] and he, no less than the Lick astronomers and M. Perrotin at Nice,[1007] observed brilliant clouds projecting beyond the terminator, or above the limb, while carried round by the planet's rotation. They seemed to float at an altitude of at least twenty miles, or about four times the height of terrestrial cirrus; but this was not wonderful, considering the low power of gravity acting upon them. Great capital was made in the journalistic interest out of these imaginary signals from intelligent Martians, desirous of opening communications with (to them) problematical terrestrial beings. Similar effects had, however, been seen before by Mr. Knobel in 1873, by M. Terby in 1888, and at the Lick Observatory in 1890; and they were discerned again with particular distinctness by Professor Hussey at Lick, August 27, 1896.[1008]

The first photograph of Mars was taken by Gould at Cordoba in 1879. Little real service in planetary delineation has, it is true, been so far rendered by the art, yet one achievement must be recorded to its credit. A set of photographs obtained by Mr. W. H. Pickering on Wilson's Peak, California, April 9, 1890, showed the southern polar cap of Mars as of moderate dimensions, but with a large dim adjacent area. Twenty-four hours later, on a corresponding set, the dim area was brilliantly white. The polar cap had become enlarged in the interim, apparently through a wide-spreading snow-fall, by the annexation of a territory equal to that of the United States. The season was towards the close of winter in Mars. Never until then had the process of glacial extension been actually (it might be said) superintended in that distant globe.

Mars was gratuitously supplied with a pair of satellites long before he was found actually to possess them. Kepler interpreted Galileo's anagram of the "triple" Saturn in this sense; they were perceived by MicromÉgas on his long voyage through space; and the Laputan astronomers had even arrived at a knowledge, curiously accurate under the circumstances, of their distances and periods. But terrestrial observers could see nothing of them until the night of August 11, 1877. The planet was then within one month of its second nearest approach to the earth during the last century; and in 1845 the Washington 26-inch refractor was not in existence.[1009] Professor Asaph Hall, accordingly, determined to turn the conjecture to account for an exhaustive inquiry into the surroundings of Mars. Keeping his glaring disc just outside the field of view, a minute attendant speck of light was "glimpsed" August 11. Bad weather, however, intervened, and it was not until the 16th that it was ascertained to be what it appeared—a satellite. On the following evening a second, still nearer to the primary, was discovered, which, by the bewildering rapidity of its passages hither and thither, produced at first the effect of quite a crowd of little moons.[1010]

Both these delicate objects have since been repeatedly observed, both in Europe and America, even with comparatively small instruments. At the opposition of 1884, indeed, the distance of the planet was too great to permit of the detection of both elsewhere than at Washington. But the Lick equatoreal showed them, July 18, 1888, when their brightness was only 0·12 its amount at the time of their discovery; so that they can now be followed for a considerable time before and after the least favourable oppositions.

The names chosen for them were taken from the Iliad, where "Deimos" and "Phobos" (Fear and Panic) are represented as the companions in battle of Ares. In several respects, they are interesting and remarkable bodies. As to size, they may be said to stand midway between meteorites and satellites. From careful photometric measures executed at Harvard in 1877 and 1879, Professor Pickering concluded their diameters to be respectively six and seven miles.[1011] This is on the assumption that they reflect the same proportion of the light incident upon them that their primary does. But it may very well be that they are less reflective, in which case they would be more extensive. The albedo of Mars is put by MÜller at 0·27; his surface, in other words, returns 27 per cent. of the rays striking it. If we put the albedo of his satellites equal to that of our moon, 0·17, their diameters will be increased from 6 and 7 to 7-1/2 and 9 miles, Phobos, the inner one, being the larger. Mr. Lowell, however, formed a considerably larger estimate of their dimensions.[1012] It is interesting to note that Deimos, according to Professor Pickering's very distinct perception, does not share the reddish tint of Mars.

Deimos completes its nearly circular revolutions in thirty hours eighteen minutes, at a distance from the surface of its ruling body of 12,500 miles; Phobos traverses an elliptical orbit[1013] in seven hours thirty-nine minutes twenty-two seconds, at a distance of only 3,760 miles. This is the only known instance of a satellite circulating faster than its primary rotates, and is a circumstance of some importance as regards theories of planetary development. To a Martian spectator the curious effect would ensue of a celestial object, seemingly exempt from the general motion of the sphere, rising in the west, setting in the east, and culminating twice, or even thrice a day; which, moreover, in latitudes above 69° north or south, would be permanently and altogether hidden by the intervening curvature of the globe.

The detection of new members of the solar system has come to be one of the most ordinary of astronomical events. Since 1846 no single year has passed without bringing its tribute of asteroidal discovery. In the last of the seventies alone, a full score of miniature planets were distinguished from the thronging stars amid which they seem to move; 1875 brought seventeen such recognitions; their number touched a minimum of one in 1881; it rose in 1882, and again in 1886, to eleven; dropped to six in 1889, and sprang up with the aid of photography to twenty-seven in 1892. That high level has since, on an average, been maintained; and on January 1, 1902, nearly 500 asteroids were recognised as revolving between the orbits of Mars and Jupiter. Of these, considerably more than one hundred are claimed by one investigator alone—Dr. Max Wolf of Heidelburg; M. Charlois of Nice comes second with 102; while among the earlier observers Palisa of Vienna contributed 86, and C. H. F. Peters of Clinton (N. Y.), whose varied and useful career terminated July 19, 1890, 52 to the grand total.

The construction by Chacornac and his successors at Paris, and more recently by Peters at Clinton, of ecliptical charts showing all stars down to the thirteenth and fourteenth magnitudes respectively, rendered the picking out of moving objects above that brightness a mere question of time and diligence. Both, however, are vastly economised by the photographic method. Tedious comparisons of the sky with charts are no longer needed for the identification of unrecorded, because simulated stars. Planetary bodies declare themselves by appearing upon the plate, not in circular, but in linear form. Their motion converts their images into trails, long or short according to the time of exposure. The first asteroid (No. 323) thus detected was by Max Wolf, December 22, 1891.[1014] Eighteen others were similarly discovered in 1892, by the same skilful operator; and ten more through Charlois's adoption at Nice of the novel plan now in exclusive use for picking up errant light-specks. Far more onerous than the task of their discovery is that of keeping them in view once discovered—of tracking out their paths, fixing their places, and calculating the disturbing effects upon them of the mighty Jovian mass. These complex operations have come to be centralised at Berlin under the superintendence of Professor Tietjen, and their results are given to the public through the medium of the Berliner Astronomisches Jahrbuch.

The cui bono? however, began to be agitated. Was it worth while to maintain a staff of astronomers for the sole purpose of keeping hold over the identity of the innumerable component particles of a cosmical ring? The prospect, indeed, of all but a select few of the asteroids being thrown back by their contemptuous captors into the sea of space seemed so imminent that Professor Watson provided by will against the dereliction of the twenty-two discovered by himself. But the fortunes of the whole family improved through the distinction obtained by one of them. On August 14, 1898, the trail of a rapidly-moving, star-like object of the eleventh magnitude imprinted itself on a plate exposed by Herr Witt at the Urania Observatory, Berlin. Its originator proved to be unique among asteroids. "Eros" is, in sober fact,

divined or imagined by Shelley.[1015] True, several of its congeners invade the Martian sphere at intervals; but the proper habitat of Eros is within that limit, although its excursions transcend it. In other words, its mean distance from the sun is about 135, as compared with the Martian distance of 141 million miles. Further, its orbit being so fortunately circumstanced as to bring it once in sixty-seven years within some 15 millions of miles of the earth, it is of extraordinary value to celestial surveyors. The calculation of its movements was much facilitated by detections, through a retrospective search,[1016] of many of its linear images among the star-dots on the Harvard plates.[1017] The little body—which can scarcely be more than twenty miles in diameter—shows peculiarities of behaviour as well as of position. Dr. von Oppolzer, in February, 1901,[1018] announced it to be extensively and rapidly variable. Once in 2 hours 38 minutes it lost about three-fourths of its light,[1019] but these fluctuations quickly diminished in range, and in the beginning of May ceased altogether.[1020] Evidently, then, they depend upon the situation of the asteroid relatively to ourselves; and, so far, events lent countenance to M. AndrÉ's eclipse hypothesis, since mutual occultations of the supposed planetary twins could only take place when the plane of their revolutions passed through the earth, and this condition would be transitory. Yet the recognition in Eros of an "Algol asteroid" seems on other grounds inadmissible;[1021] nor until the phenomenon is conspicuously renewed—as it probably will be at the opposition of 1903—can there be much hope of finding its appropriate rationale.

The crowd of orbits disclosed by asteroidal detections invites attentive study. D'Arrest remarked in 1851,[1022] when only thirteen minor planets were known, that supposing their paths to be represented by solid hoops, not one of the thirteen could be lifted from its place without bringing the others with it. The complexity of interwoven tracks thus illustrated has grown almost in the numerical proportion of discovery. Yet no two actually intersect, because no two lie exactly in the same plane, so that the chances of collision are at present nil. There is only one case, indeed, in which it seems to be eventually possible. M. Lespiault has pointed out that the curves traversed by "FidÉs" and "MaÏa" approach so closely that a time may arrive when the bodies in question will either coalesce or unite to form a binary system.[1023]

The maze threaded by the 500 asteroids contrasts singularly with the harmoniously ordered and rhythmically separated orbits of the larger planets. Yet the seeming confusion is not without a plan.

The established rules of our system are far from being totally disregarded by its minor members. The orbit of Pallas, with its inclination of 34° 42', touches the limit of departure from the ecliptic level; the average obliquity of the asteroidal paths is somewhat less than that of the sun's equator;[1024] their mean eccentricity is below that of the curve traced out by Mercury, and all without exception are pursued in the planetary direction—from west to east.

The zone in which these small bodies travel is about three times as wide as the interval separating the earth from the sun. It extends perilously near to Jupiter, and dovetails into the sphere of Mars.

Their distribution is very unequal. They are most densely congregated about the place where a single planet ought, by Bode's Law, to revolve; it may indeed be said that only stragglers from the main body are found more than fifty million miles within or without a mean distance from the sun 2·8 times that of the earth. Significant gaps, too, occur where some force prohibitive of their presence would seem to be at work. The probable nature of that force was suggested by the late Professor Kirkwood, first in 1866, when the number of known asteroids was only eighty-eight, and again with more confidence in 1876, from the study of a list then run up to 172.[1025] It appears that these bare spaces are found just where a revolving body would have a period connected by a simple relation with that of Jupiter. It would perform two or three circuits to his one, five to his two, nine to his five, and so on. Kirkwood's inference was that the gaps in question were cleared of asteroids by the attractive influence of Jupiter. For disturbances recurring time after time—owing to commensurability of periods—nearly at the same part of the orbit, would have accumulated until the shape of that orbit was notably changed. The body thus displaced would have come in contact with other cosmical particles of the same family with itself—then, it may be assumed, more evenly scattered than now—would have coalesced with them, and permanently left its original track. In this way the regions of maximum perturbation would gradually have become denuded of their occupants.

We can scarcely doubt that this law of commensurability has largely influenced the present distribution of the asteroids. But its effects must have been produced while they were still in an unformed, perhaps a nebular condition. In a system giving room for considerable modification through disturbance, the recurrence of conjunctions with a dominating mass at the same orbital point need not involve instability.[1026] On the whole, the correspondence of facts with Kirkwood's hypothesis has not been impaired by their more copious collection.[1027] Some chasms of secondary importance have indeed been bridged; but the principal stand out more conspicuously through the denser scattering of orbits near their margins. Nor is it doubtful that the influence of Jupiter in some way produced them. M. de Freycinet's study of the problem they present[1028] has, however, led him to the conclusion that they existed ab origine, thus testifying rather to the preventive than to the perturbing power of the giant planet.

The existence, too, of numerous asteroidal pairs travelling in approximately coincident tracks, must date from a remote antiquity. They result, Professor Kirkwood[1029] believed, from the divellent action of Jupiter upon embryo pigmy planets, just as comets moving in pursuit of one another are a consequence of the sundering influence of the sun.

Leverrier fixed, in 1853,[1030] one-fourth of the earth's mass as the outside limit for the combined masses of all the bodies circulating between Mars and Jupiter; but it is far from probable that this maximum is at all nearly approached. M. Berberich[1031] held that the moon would more than outweigh the whole of them, a million of the lesser bodies shining like stars of the twelfth magnitude being needed, according to his judgment, to constitute her mass. And M. Niesten estimated that the whole of the 216 asteroids discovered up to August, 1880, amounted in volume to only 1/4000th of our globe,[1032] and we may safely add—since they are tolerably certain to be lighter, bulk for bulk, than the earth—that their proportionate mass is smaller still. A fairly concordant result was published in 1895 by Mr. B. M. Roszel.[1033] He found that the lunar globe probably contains forty times, the terrestrial globe 3,240 times the quantity of matter parcelled out among the first 311 minor planets. The actual size of a few of them may now be said to be known. Professor Pickering, from determinations of light-intensity, assigned to Vesta a diameter of 319 miles, to Pallas 167, to Juno 94, down to twelve and fourteen for the smaller members of the group.[1034] An albedo equal to that of Mars was assumed as the basis of the calculation. Moreover, Professor G. MÜller[1035] of Potsdam examined photometrically the phases of seven among them, of which four—namely, Vesta, Iris, Massalia, and Amphitrite—were found to conform precisely to the behaviour of Mars as regards light-change from position, while Ceres, Pallas, and Irene varied after the manner of the moon and Mercury. The first group were hence inferred to resemble Mars in physical constitution, nature of atmosphere, and reflective capacity; the second to be moon-like bodies.

Finally, Professor Barnard, directly measuring with the Yerkes refractor the minute discs presented by the original quartette, obtained the following authentic data concerning them:[1036] Diameter of Ceres, 477 miles, albedo = 0·18; diameter of Pallas, 304 miles, albedo = 0·23; diameter of Vesta, 239 miles, albedo = 0·74; diameter of Juno, 120 miles, albedo = 0·45. Thus, the rank of premier asteroid proves to belong to Ceres, and to have been erroneously assigned to Vesta in consequence of its deceptive brilliancy. What kind of surface this indicates, it is hard to say. The dazzling whiteness of snow can hardly be attributed to bare rock; yet the dynamical theory of gases—as Dr. Johnstone Stoney pointed out in 1867[1037]—prohibits the supposition that bodies of insignificant gravitative power can possess aerial envelopes. Even our moon, it is calculated, could not permanently hold back the particles of oxygen, nitrogen, or water-gas from escaping into infinite space; still less, a globe one thousand times smaller. Vogel's suspicion of an air-line in the spectrum of Vesta[1038] has, accordingly, not been confirmed.

Crossing the zone of asteroids on our journey outward from the sun, we meet with a group of bodies widely different from the "inferior" or terrestrial planets. Their gigantic size, low specific gravity, and rapid rotation, obviously from the first threw the "superior" planets into a class apart; and modern research has added qualities still more significant of a dissimilar physical constitution. Jupiter, a huge globe 86,000 miles in diameter, stands pre-eminent among them. He is, however, only primus inter pares; all the wider inferences regarding his condition may be extended, with little risk of error, to his fellows; and inferences in his case rest on surer grounds than in the case of the others, from the advantages offered for telescopic scrutiny by his comparative nearness.

Now the characteristic modern discovery concerning Jupiter is that he is a body midway between the solar and terrestrial stages of cosmical existence—a decaying sun or a developing earth, as we choose to put it—whose vast unexpended stores of internal heat are mainly, if not solely, efficient in producing the interior agitations betrayed by the changing features of his visible disc. This view, impressed upon modern readers by Mr. Proctor's popular works, was anticipated in the last century. Buffon wrote in his Époques de la Nature (1778):[1039]—"La surface de Jupiter est, comme l'on sait, sujette À des changemens sensibles, qui semblent indiquer que cette grosse planÈte est encore dans un État d'inconstance et de bouillonnement."

Primitive incandescence, attendant, in his fantastic view, on planetary origin by cometary impacts with the sun, combined, he concluded, with vast bulk to bring about this result. Jupiter has not yet had time to cool. Kant thought similarly in 1785;[1040] but the idea did not commend itself to the astronomers of the time, and dropped out of sight until Mr. Nasmyth arrived at it afresh in 1853.[1041] Even still, however, terrestrial analogies held their ground. The dark belts running parallel to the equator, first seen at Naples in 1630, continued to be associated—as Herschel had associated them in 1781—with Jovian trade-winds, in raising which the deficient power of the sun was supposed to be compensated by added swiftness of rotation. But opinion was not permitted to halt here.

In 1860 G. P. Bond of Cambridge (U.S.) derived some remarkable indications from experiments on the light of Jupiter.[1042] They showed that fourteen times more of the photographic rays striking it are reflected by the planet than by our moon, and that, unlike the moon, which sends its densest rays from the margin, Jupiter is brightest near the centre. But the most perplexing part of his results was that Jupiter actually seemed to give out more light than he received. Bond, however, rightly considered his data too uncertain for the support of so bold an assumption as that of original luminosity, and, even if the presence of native light were proved, thought that it might emanate from auroral clouds of the terrestrial kind. The conception of a sun-like planet was still a remote, and seemed an extravagant one.

Only since it was adopted and enforced by ZÖllner in 1865,[1043] can it be regarded as permanently acquired to science. The rapid changes in the cloud-belts both of Jupiter and Saturn, he remarked, attest a high internal temperature. For we know that all atmospheric movements on the earth are sun-heat transformed into motion. But sun-heat at the distance of Jupiter possesses but 1/27, at that of Saturn 1/100 of its force here. The large amount of energy, then, obviously exerted in those remote firmaments must have some other source, to be found nowhere else than in their own active and all-pervading fires, not yet banked in with a thick solid crust.

The same acute investigator dwelt, in 1871,[1044] on the similarity between the modes of rotation of the great planets and of the sun, applying the same principles of explanation to each case. The fact of this similarity is undoubted. Cassini[1045] and SchrÖter both noticed that markings on Jupiter travelled quicker the nearer they were to his equator; and Cassini even hinted at their possible assimilation to sun-spots.[1046] It is now well ascertained that, as a rule (not without exceptions), equatorial spots give a period some 5-1/2 minutes shorter than those in latitudes of about 30°. But, as Mr. Denning has pointed out,[1047] no single period will satisfy the observations either of different markings at the same epoch, or of the same markings at different epochs. Accelerations and retardations, depending upon processes of growth or change, take place in very much the same kind of way as in solar maculÆ, inevitably suggesting similarity of origin.

The interesting query as to Jupiter's surface incandescence has been studied since Bond's time with the aid of all the appliances furnished to physical inquirers by modern inventiveness, yet without bringing to it a categorical reply. ZÖllner in 1865, MÜller in 1893, estimated his albedo at 0·62 and 0·75 respectively, that of fresh-fallen snow being 0·78, and of white paper 0·70.[1048] But the disc of Jupiter is by no means purely white. The general ground is tinged with ochre; the polar zones are leaden or fawn coloured; large spaces are at times stained or suffused with chocolate-browns and rosy hues. It is occasionally seen ruled from pole to pole with dusky bars, and is never wholly free from obscure markings. The reflection, then, by it, as a whole, of about 70 per cent. of the rays impinging upon it, might well suggest some original reinforcement.

Nevertheless, the spectroscope gives little countenance to the supposition of any considerable permanent light-emission. The spectrum of Jupiter, as examined by Huggins, 1862-64, and by Vogel, 1871-73, shows the familiar Fraunhofer rays belonging to reflected sunlight. But it also shows lines of native absorption. Some of these are identical with those produced by the action of our own atmosphere, especially one or more groups due to aqueous vapours; others are of unknown origin; and it is remarkable that one among the latter—a strong band in the red—agrees in position with a dark line in the spectra of some ruddy stars.[1049] There is, besides, a general absorption of blue rays, intensified—as Le Sueur observed at Melbourne in 1869[1050]—in the dusky markings, evidently through an increase of depth in the atmospheric strata traversed by the light proceeding from them.

All these observations, however (setting aside the stellar line as of doubtful significance), point to a cool planetary atmosphere. One spectrograph, it is true, taken by Dr. Henry Draper, September 27, 1879,[1051] seemed to attest the action of intrinsic light; but the peculiarity was referred by Dr. Vogel, with convincing clearness, to a flaw in the film.[1052] So far, then, native emissions from any part of Jupiter's diversified surface have not been detected; and, indeed, the blackness of the shadows cast by his satellites on his disc sufficiently proves that he sends out virtually none but reflected light.[1053] This conclusion, however, by no means invalidates that of his high internal temperature.

The curious phenomena attending Jovian satellite-transits may be explained, partly as effects of contrast, partly as due to temporary obscurations of the small discs projected on the large disc of Jupiter. At their first entry upon its marginal parts, which are several times less luminous than those near the centre, they invariably show as bright spots, then usually vanish as the background gains lustre, to reappear, after crossing the disc, thrown into relief, as before, against the dusky limb. But instances are not rare, more especially of the third and fourth satellites standing out, during the entire middle part of their course, in such inky darkness as to be mistaken for their own shadows. The earliest witness of a "black transit" was Cassini, September 2, 1665; RÖmer in 1677, and Maraldi in 1707 and 1713, made similar observations, which have been multiplied in recent years. In some cases the process of darkening has been visibly attended by the formation, or emergence into view, of spots on the transiting body, as noted by the two Bonds at Harvard, March 18, 1848.[1054] The third satellite was seen by Dawes, half dark, half bright, when crossing Jupiter's disc, August 21, 1867;[1055] one-third dark by Davidson of California, January 15, 1884, under the same circumstances;[1056] and unmistakably spotted, both on and off the planet, by SchrÖter, Secchi, Dawes, and Lassell.

The first satellite sometimes looks dusky, but never absolutely black, in travelling over the disc of Jupiter. The second appears uniformly white—a circumstance attributed by Dr. Spitta[1057] to its high albedo. The singularly different aspects, even during successive transits, of the third and fourth satellites, are connected by Professor Holden[1058] with the varied luminosity of the segments of the planetary surface they are projected upon, and W. H. Pickering inclines to the same opinion; but fluctuations in their own brightness[1059] may be a concurrent cause. Herschel concluded in 1797 that, like our moon, they always turn the same face towards their primary, and as regards the outer satellite, Engelmann's researches in 1871, and C. E. Burton's in 1873, made this almost certain; while both for the third and fourth Jovian moons it was completely assured by W. H. Pickering's and A. E. Douglass's observations at Arequipa in 1892,[1060] and at Flagstaff in 1894-95.[1061] Strangely enough, however, the interior members of the system have preserved a relatively swift rotation, notwithstanding the enormous checking influence upon it of Jove-raised tides.

All the satellites are stated, on good authority, to be more or less egg-shaped. On September 8, 1890, Barnard saw the first elongated and bisected by a bright equatorial belt, during one of its dark transits;[1062] and his observation, repeated August 3, 1891, was completely verified by Schaeberle and Campbell, who ascertained, moreover, that the longer axis of the prolate body was directed towards Jupiter's centre.[1063] The ellipticity of its companions was determined by Pickering and Douglass; indeed, that of No. 3 had long previously been noticed by Secchi.[1064] No. 3 also shows equatorial stripes, perceived in 1891 by Schaeberle and Campbell,[1065] and evident later to Pickering and Douglass;[1066] nor need we hesitate to admit as authentic their records of similar, though less conspicuous markings on the other satellites. A constitution analogous to that of Jupiter himself was thus unexpectedly suggested; and Vogel's detection of lines—or traces of lines—in their spectra, agreeing with absorption-rays derived from their primary, lends support to the conjecture that they possess gaseous envelopes similar to his.

The system of Jupiter, as it was discovered by Galileo, and investigated by Laplace, appeared in its outward aspect so symmetrical, and displayed in its inner mechanism such harmonious dynamical relations, that it might well have been deemed complete. Nevertheless, a new member has been added to it. Near midnight on September 9, 1892, Professor Barnard discerned with the Lick 36-inch "a tiny speck of light," closely following the planet.[1067] He instantly divined its nature, watched its hurried disappearance in the adjacent glare, and made sure of the reality of his discovery on the ensuing night. It was a delicate business throughout, the Liliputian luminary subsiding into invisibility before the slightest glint of Jovian light, and tarrying, only for brief intervals, far enough from the disc to admit of its exclusion by means of an occulting plate. The new satellite is estimated to be of the thirteenth stellar magnitude, and, if equally reflective of light with its next neighbour, Io (satellite No. 1), its diameter must be about one hundred miles. It revolves at a distance of 112,500 miles from Jupiter's centre, and of 68,000 from his bulging equatorial surface. Its period of 11h. 57m. 23s. is just two hours longer than Jupiter's period of rotation, so that Phobos still remains a unique example of a secondary body revolving faster than its primary rotates. Jupiter's innermost moon conforms in its motions strictly, indeed inevitably, to the plane of his equatorial protuberance, following, however, a sensibly elliptical path the major axis of which is in rapid revolution.[1068] Its very insignificance raises the suspicion that it may not prove solitary. Possibly it belongs to a zone peopled by asteroidal satellites. More than fifteen thousand such small bodies could be furnished out of the materials of a single full-sized satellite spoiled in the making. But we must be content for the present to register the fact without seeking to penetrate the meaning of its existence. Very high and very fine telescopic power is needed for its perception. Outside the United States, it has been very little observed. The only instruments in this country successfully employed for its detection are, we believe, Dr. Common's 5-foot reflector and Mr. Newall's 25-inch refractor.

In the course of his observations on Jupiter at Brussels in 1878, M. Niesten was struck with a rosy cloud attached to a whitish zone beneath the dark southern equatorial band.[1069] Its size was enormous. At the distance of Jupiter, its measured dimensions of 13' by 3' implied a real extension in longitude of 30,000, in latitude of something short of 7,000 miles. The earliest record of its appearance seems to be by Professor Pritchett, director of the Morrison Observatory (U.S.), who figured and described it July 9, 1878.[1070] It was again delineated August 9, by Tempel at Florence.[1071] In the following year it attracted the wonder and attention of almost every possessor of a telescope. Its colour had by that time deepened into a full brick-red, and was set off by contrast with a white equatorial spot of unusual brilliancy. During three ensuing years these remarkable objects continued to offer a visible and striking illustration of the compound nature of the planet's rotation. The red spot completed a circuit in nine hours fifty-five minutes thirty-six seconds; the white spot in about five and a half minutes less. Their relative motion was thus no less than 260 miles an hour, bringing them together in the same meridian at intervals of forty-four days ten hours forty-two minutes. Neither, however, preserved continuously the same uniform rate of travel. The period of each had lengthened by some seconds in 1883, while sudden displacements, associated with the recovery of lustre after recurrent fadings, were observed in the position of the white spot,[1072] recalling the leap forward of a reviving sun-spot. Just the opposite effect attended the rekindling of the companion object. While semi-extinct, in 1882-84, it lost little motion; but a fresh access of retardation was observed by Professor Young[1073] in connection with its brightening in 1886. This suggests very strongly that the red spot is fed from below. A shining aureola of "faculÆ," described by Bredichin at Moscow, and by Lohse at Potsdam, as encircling it in September, 1879,[1074] was held to strengthen the solar analogy.

The conspicuous visibility of this astonishing object lasted three years. When the planet returned to opposition in 1882-83, it had faded so considerably that RiccÒ's uncertain glimpse of it at Palermo, May 31, 1883, was expected to be the last. It had, nevertheless, begun to recover in December, and presented to Mr. Denning in the beginning of 1886 much the same aspect as in October, 1882.[1075] Observed by him in an intermediate stage, February 25, 1885, when "a mere skeleton of its former self," it bore a striking likeness to an "elliptical ring" descried in the same latitude by Mr. Gledhill in 1869-70. This, indeed, might be called the preliminary sketch for the famous object brought to perfection ten years later, but which Mr. H. C. Russell of Sydney saw and drew still unfinished in June, 1876,[1076] before it had separated from its matrix, the dusky south tropical belt. In earlier times, too, a marking "at once fixed and transient" had been repeatedly perceived attached to the southernmost of the central belts. It gave Cassini in 1665 a rotation-period of nine hours fifty-six minutes,[1077] reappeared and vanished eight times during the next forty-three years, and was last seen by Maraldi in 1713. It was, however, very much smaller than the recent object, and showed no unusual colour.[1078]

The assiduous observations made on the "Great Red Spot" by Mr. Denning at Bristol and by Professor Hough at Chicago afforded grounds only for negative conclusions as to its nature. It certainly did not represent the outpourings of a Jovian volcano; it was in no sense attached to the Jovian soil—if the phrase have any application to that planet; it was not a mere disclosure of a glowing mass elsewhere seethed over by rolling vapours. It was, indeed, certainly not self-luminous, a satellite projected upon it in transit having been seen to show as bright as upon the dusky equatorial bands. A fundamental objection to all three hypotheses is that the rotation of the spot was variable. It did not then ride at anchor, but floated free. Some held that its surface was depressed below the average cloud-level, and that the cavity was filled with vapours. Professor Wilson, on the other hand, observing with the 16-inch equatorial of the Goodsell Observatory in Minnesota, received a persistent impression of the object "being at a higher level than the other markings."[1079] A crucial experiment on this point was proposed by Mr. Stanley Williams in 1890.[1080] A dark spot moving faster along the same parallel was timed to overtake the red spot towards the end of July. A unique opportunity hence appeared to be at hand of determining the relative vertical depths of the two formations, one of which must inevitably, it was thought, pass above the other. No forecast included a third alternative, which was nevertheless adopted by the dark spot. It evaded the obstacle in its path by skirting round its southern edge.[1081] Nothing, then, was gained by the conjunction, beyond an additional proof of the singular repellent influence exerted by the red spot over the markings in its vicinity. It has, for example, gradually carved out a deep bay for its accommodation in the gray belt just north of it. The effect was not at first steadily present. A premonitory excavation was drawn by Schwabe at Dessau, September 5, 1831, and again by Trouvelot, Barnard, and Elvins in 1879; yet there was no sign of it in the following year. Its development can be traced in Dr. Boeddicker's beautiful delineations of Jupiter, made with the Parsonstown 3-foot reflector, from 1881 to 1886.[1082] They record the belt as straight in 1881, but as strongly indented from January, 1883; and the cavity now promises to outlast the spot. So long as it survives, however, the forces at work in the spot can have lost little of their activity. For it must be remembered that the belt has a shorter rotation-period than the red spot, which, accordingly (as Mr. Elvins of Toronto has pointed out), breasts and diverts, by its interior energy, a current of flowing matter, ever ready to fill up its natural bed, and override the barrier of obstruction.

The famous spot was described by Keeler in 1889, as "of a pale pink colour, slightly lighter in the middle. Its outline was a fairly true ellipse, framed in by bright white clouds."[1083] The fading continuously in progress from 1887 was temporarily interrupted in 1891. The revival, indeed, was brief. Professor Barnard wrote in August, 1892: "The great red spot is still visible, but it has just passed through a crisis that seemingly threatened its very existence. For the past month it has been all but impossible to catch the feeblest trace of the spot, though the ever-persistent bay in the equatorial belt close north of it, and which has been so intimately connected with the history of the red spot, has been as conspicuous as ever. It is now, however, possible to detect traces of the entire spot. An obscuring medium seems to have been passing over it, and has now drifted somewhat preceding the spot."[1084]

The object is now always inconspicuous, and often practically invisible, and may be said to float passively in the environing medium.[1085] Yet there are sparks beneath the ashes. A rosy tinge faintly suffused it in April, 1900,[1086] and its absolute end may still be remote.

The extreme complexity of the planet's surface-movements has been strikingly evinced by Mr. Stanley Williams's detailed investigations. He enumerated in 1896[1087] nine principal currents, all flowing parallel to the equator, but unsymmetrically placed north and south of it, and showing scant signs of conformity to the solar rule of retardation with increase of latitude. The linear rate of the planet's equatorial rotation was spectroscopically determined by BÉlopolsky and Deslandres in 1895. Both found it to fall short of the calculated speed, whence an enlargement, by self-refraction, of the apparent disc was inferred.[1088]

Jupiter was systematically photographed with the Lick 36-inch telescope during the oppositions of 1890, 1891, and 1892, the image thrown on the plates (after eightfold direct enlargement) being one inch in diameter. Mr. Stanley Williams's measurements and discussion of the set for 1891 showed the high value of the materials thus collected, although much more minute details can be seen than can at present be photographed. The red spot shows as "very distinctly annular" in several of these pictures.[1089] Recently, the planet has been portrayed by Deslandres with the 62-foot Meudon refractor.[1090] The extreme actinic feebleness of the equatorial bands was strikingly apparent on his plates.

In 1870, Mr. Ranyard[1091]—whose death, December 14, 1894, was a serious loss to astronomy—acting upon an earlier suggestion of Sir William Huggins, collected records of unusual appearances on the disc of Jupiter, with a view to investigate the question of their recurrence at regular intervals. He concluded that the development of the deeper tinges of colour, and of the equatorial "port-hole" markings girdling the globe in regular alternations of bright and dusky, agreed, so far as could be ascertained, with epochs of sun-spot maximum. The further inquiries of Dr. Lohse at Bothkamp in 1873[1092] went to strengthen the coincidence, which had been anticipated À priori by ZÖllner in 1871.[1093] Moreover, separate and distinct evidence was alleged by Mr. Denning in 1899 of decennial outbreaks of disturbance in north temperate regions.[1094] It may, indeed, be taken for granted that what Hahn terms the universal pulse of the solar system[1095] affects the vicissitudes of Jupiter; but the law of those vicissitudes is far from being so obviously subordinate to the rhythmical flow of central disturbance as are certain terrestrial phenomena. The great planet, being in fact himself a "semi-sun," may be regarded as an originator, no less than a recipient, of agitating influences, the combined effects of which may well appear insubordinate to any obvious law.

It is likely that Saturn is in a still earlier stage of planetary development than Jupiter. He is the lightest for his size of all the planets. In fact, he would float in water. And since his density is shown, by the amount of his equatorial bulging, to increase centrally,[1096] it follows that his superficial materials must be of a specific gravity so low as to be inconsistent, on any probable supposition, with the solid or liquid states. Moreover, the chief arguments in favour of the high temperature of Jupiter, apply, with increased force, to Saturn; so that it may be concluded, without much risk of error, that a large proportion of his bulky globe, 73,000 miles in diameter, is composed of heated vapours, kept in active and agitated circulation by the process of cooling.

His unique set of appendages has, since the middle of the last century, formed the subject of searching and fruitful inquiries, both theoretical and telescopic. The mechanical problem of the stability of Saturn's rings was left by Laplace in a very unsatisfactory condition. Considering them as rotating solid bodies, he pointed out that they could not maintain their position unless their weight were in some way unsymmetrically distributed; but made no attempt to determine the kind or amount of irregularity needed to secure this end. Some observations by Herschel gave astronomers an excuse for taking for granted the fulfilment of the condition thus vaguely postulated; and the question remained in abeyance until once more brought prominently forward by the discovery of the dusky ring in 1850.

The younger Bond led the way, among modern observers, in denying the solidity of the structure. The fluctuations in its aspect were, he asserted in 1851,[1097] inconsistent with such a hypothesis. The fine dark lines of division, frequently detected in both bright rings, and as frequently relapsing into imperceptibility, were due, in his opinion, to the real nobility of their particles, and indicated a fluid formation. Professor Benjamin Peirce of Harvard University immediately followed with a demonstration, on abstract grounds, of their non-solidity.[1098] Streams of some fluid denser than water were, he maintained, the physical reality giving rise to the anomalous appearance first disclosed by Galileo's telescope.

The mechanism of Saturn's rings, proposed as the subject of the Adams Prize, was dealt with by James Clerk Maxwell in 1857. His investigation forms the groundwork of all that is at present known in the matter. Its upshot was to show that neither solid nor fluid rings could continue to exist, and that the only possible composition of the system was by an aggregated multitude of unconnected particles, each revolving independently in a period corresponding to its distance from the planet.[1099] This idea of a satellite-formation had been, remarkably enough, several times entertained and lost sight of. It was first put forward by Roberval in the seventeenth century, again by Jacques Cassini in 1715, and with perfect definiteness by Wright of Durham in 1750.[1100] Little heed, however, was taken of these casual anticipations of a truth which reappeared, a virtual novelty, as the legitimate outcome of the most refined modern methods.

The details of telescopic observation accord, on the whole, admirably with this hypothesis. The displacements or disappearance of secondary dividing-lines—the singular striated appearance, first remarked by Short in the eighteenth century, last by Perrotin and Lockyer at Nice, March 18, 1884[1101]—show the effects of waves of disturbance traversing a moving mass of gravitating particles;[1102] the broken and changing line of the planet's shadow on the ring gives evidence of variety in the planes of the orbits described by those particles. The whole ring-system, too, appears to be somewhat elliptical.[1103]

The satellite-theory has derived unlooked-for support from photometric inquiries. Professor Seeliger pointed out in 1888[1104] that the unvarying brilliancy of the outer rings under all angles of illumination, from 0° to 30°, can be explained from no other point of view. Nor does the constitution of the obscure inner ring offer any difficulty. For it is doubtless formed of similar small bodies to those aggregated in the lucid members of the system, only much more thinly strewn, and reflecting, consequently, much less light. It is not, indeed, at first easy to see why these sparser flights should show as a dense dark shading on the body of Saturn. Yet this is invariably the case. The objection has been urged by Professor Hastings of Baltimore. The brightest parts of these appendages, he remarked,[1105] are more lustrous than the globe they encircle; but if the inner ring consists of identical materials, possessing presumably an equal reflective capacity, the mere fact of their scanty distribution would not cause them to show as dark against the same globe. Professor Seeliger, however, replied[1106] that the darkening is due to the never-ending swarms of their separate shadows transiting the planet's disc. Sunlight is not, indeed, wholly excluded. Many rays come and go between the open ranks of the meteorites. For the dusky ring is transparent. The planet it encloses shows through it, as if veiled with a strip of crape. A beautiful illustration of its quality in this respect was derived by Professor Barnard from an eclipse of Japetus, November 1, 1889.[1107] The eighth moon remained steadily visible during its passage through the shadow of the inner ring, but with a progressive loss of lustre in approaching its bright neighbour. There was no breach of continuity. The satellite met no gap, corresponding to that between the dusky ring and the body of Saturn, through which it could shine with undiminished light, but was slowly lost sight of as it plunged into deeper and deeper gloom. The important facts were thus established, that the brilliant and obscure rings merge into each other, and that the latter thins out towards the Saturnian globe.

The meteoric constitution of these appendages was beautifully demonstrated in 1895 by Professor Keeler,[1108] then director of the Alleghany Observatory, Pittsburgh. From spectrographs taken with the slit adjusted to coincidence with the equatorial plane of the system, he determined the comparative radial velocities of its different parts. And these supply a crucial test of Clerk Maxwell's theory. For if the rings were solid, the swiftest rates of rotation should be at their outer edges, corresponding to wider circles described in the same period; while, if they are pulverulent, the inverse relation must hold good. This proved to be actually the case. The motion slowed off outward, in agreement with the diminishing speed of particles travelling freely, each in its own orbit. Keeler's result was promptly confirmed by Campbell,[1109] as well as by Deslandres and BÉlopolsky.

A question of singular interest, and one which we cannot refrain from putting to ourselves, is—whether we see in the rings of Saturn a finished structure, destined to play a permanent part in the economy of the system; or whether they represent merely a stage in the process of development out of the chaotic state in which it is impossible to doubt that the materials of all planets were originally merged. M. Otto Struve attempted to give a definite answer to this important query.

A study of early and later records of observations disclosed to him, in 1851, an apparent progressive approach of the inner edge of the bright ring to the planet. The rate of approach he estimated at about fifty-seven English miles a year, or 11,000 miles during the 194 years elapsed since the time of Huygens.[1110] Were it to continue, a collapse of the system must be far advanced within three centuries. But was the change real or illusory—a plausible, but deceptive inference from insecure data? M. Struve resolved to put it to the test. A set of elaborately careful micrometrical measures of the dimensions of Saturn's rings, executed by himself at Pulkowa in the autumn of 1851, was provided as a standard of future comparison; and he was enabled to renew them, under closely similar circumstances, in 1882.[1111] But the expected diminution of the space between Saturn's globe and his rings had not taken place. A slight extension in the width of the system, both outward and inward, was indeed, hinted at; and it is worth notice that just such a separation of the rings was indicated by Clerk Maxwell's theory, so that there is an À priori likelihood of its being in progress. Yet Hall's measures in 1884-87[1112] failed to supply evidence of alteration with time; and Barnard's, executed at Lick in 1894-95,[1113] showed no sensible divergence from them. Hence, much weight cannot be laid upon Huygens's drawings and descriptions, which had been held to prove conclusively a partial filling up, since 1657, of the interval between the ring and the planet.[1114]

The rings of Saturn replace, in Professor G. H. Darwin's view,[1115] an abortive satellite, scattered by tidal action into annular form. For they lie closer to the planet than is consistent with the integrity of a revolving body of reasonable bulk. The limit of possible existence for such a mass was fixed by Roche of Montpellier, in 1848,[1116] at 2·44 mean radii of its primary; while the outer edge of the ring-system is distant 2·38 radii of Saturn from his centre. The virtual discovery of its pulverulent condition dates, then, according to Professor Darwin, from 1848. He conjectures that the appendage will eventually disappear, partly through the dispersal of its constituent particles inward, and their subsidence upon the planet's surface, partly by their dispersal outward, to a region beyond "Roche's limit," where coalescence might proceed unhindered by the strain of unequal attractions. One modest satellite, revolving inside Mimas, would then be all that was left of the singular appurtenances we now contemplate with admiration.

There seems reason to admit that Kirkwood's law of commensurability has had some effect in bringing about the present distribution of the matter composing them. Here the influential bodies are Saturn's moons, while the divisions and boundaries of the rings represent the spaces where their disturbing action conspires to eliminate revolving particles. Kirkwood, in fact, showed, in 1867,[1117] that a body circulating in the chasm between the bright rings known as "Cassini's division," would have a period nearly commensurable with those of four out of the eight moons; and Meyer of Geneva subsequently calculated all such combinations, with the result of bringing out coincidences between regions of maximum perturbation and the limiting and dividing lines of the system.[1118] This is in itself a strong confirmation of the view that the rings are made up of independently revolving small bodies.

On December 7, 1876, Professor Asaph Hall discovered at Washington a bright equatorial spot on Saturn, which he followed and measured through above sixty rotations, each performed in ten hours fourteen minutes twenty-four seconds.[1119] This, he was careful to add, represented the period, not necessarily of the planet, but only of the individual spot. The only previous determination of Saturn's axial movement (setting aside some insecure estimates by SchrÖter) was Herschel's in 1794, giving a period of ten hours sixteen minutes. The substantial accuracy of Hall's result was verified by Mr. Denning in 1891.[1120] In May and June of that year, ten vague bright markings near the equator were watched by Mr. Stanley Williams, who derived from them a rotation period only two seconds shorter than that determined at Washington. Nevertheless, similarly placed spots gave in 1892 and 1893 notably quicker rates;[1121] so that the task of timing the general drift of the Saturnian surface by the displacements of such objects is hampered, to an indefinite extent, by their individual proper motions.

Saturn's outermost satellite, Japetus, is markedly variable—so variable that it sends us, when brightest, just 4-1/2 times as much light as when faintest. Moreover, its fluctuations depend upon its orbital position in such a way as to make it a conspicuous telescopic object when west, a scarcely discernible one when east of the planet. Herschel's inference[1122] of a partially obscured globe turning always the same face towards its primary seems the only admissible one, and is confirmed by Pickering's measurements of the varying intensity of its light. He remarked further that the dusky and brilliant hemispheres must be so posited as to divide the disc, viewed from Saturn, into nearly equal parts; so that this Saturnian moon, even when "full," appears very imperfectly illuminated over one-half of its surface.[1123]

ZÖllner estimated the albedo of Saturn at 0·51, MÜller at 0·88, a value impossibly high, considering that the spectrum includes no vestige of original emissions. Closely similar to that of Jupiter, it shows the distinctive dark line in the red (wave-length 618), which we may call the "red-star line"; and Janssen, from the summit of Etna in 1867[1124] found traces in it of aqueous absorption. The light from the ring appears to be pure reflected sunshine unmodified by original atmospheric action.[1125]

Uranus, when favourably situated, can easily be seen with the naked eye as a star between the fifth and sixth magnitudes. There is indeed, some reason to suppose that he had been detected as a wandering orb by savage "watchers of the skies" in the Pacific long before he swam into Herschel's ken. Nevertheless, inquiries into his physical habitudes are still in an early stage. They are exceedingly difficult of execution, even with the best and largest modern telescopes; and their results remain clouded with uncertainty.

It will be remembered that Uranus presents the unusual spectacle of a system of satellites travelling nearly at right angles to the plane of the ecliptic. The existence of this anomaly gives a special interest to investigations of his axial movement, which might be presumed, from the analogy of the other planets, to be executed in the same tilted plane. Yet this is far from being certainly the case.

Mr. Buffham in 1870-72 caught traces of bright markings on the Uranian disc, doubtfully suggesting a rotation in about twelve hours in a plane not coincident with that in which his satellites circulate.[1126] Dusky bands resembling those of Jupiter, but very faint, were barely perceptible to Professor Young at Princeton in 1883. Yet, though almost necessarily inferred to be equatorial, they made a considerable angle with the trend of the satellites' orbits.[1127] More distinctly by the brothers Henry, with the aid of their fine refractor, two gray parallel rulings, separated by a brilliant zone, were discerned every clear night at Paris from January to June, 1884.[1128] What were taken to be the polar regions appeared comparatively dusky. The direction of the equatorial rulings (for so we may safely call them) made an angle of 40° with the satellites' line of travel. Similar observations were made at Nice by MM. Perrotin and Thollon, March to June, 1884, a lucid spot near the equator, in addition, indicating rotation in a period of about ten hours.[1129] The discrepancy was, however, considerably reduced by Perrotin's study of the planet in 1889 with the new 30-inch equatoreal.[1130] The dark bands, thus viewed to better advantage than in 1884, appeared to deviate no more than 10° from the satellites' orbit-plane. No definitive results, on the other hand, were derived by Professors Holden, Schaeberle, and Keeler from their observations of Uranus in 1889-90 with the potent instrument on Mount Hamilton. Shadings, it is true, were almost always, though faintly, seen; but they appeared under an anomalous, possibly an illusory aspect. They consisted, not of parallel, but of forked bands.[1131]

Measurements of the little sea-green disc which represents to us the massive bulk of Uranus, by Young, Schiaparelli,[1132] Safarik, H. C. Wilson[1133] and Perrotin, prove it to be quite distinctly bulged. The compression at once caught Barnard's trained eye in 1894,[1134] when he undertook at Lick a micrometrical investigation of the system; and he was surprised to perceive that the major axis of the elliptical surface made an angle of about 28° with the line of travel pursued by the satellites. Nothing more can be learned on this curious subject for some years, since the pole of the planet is just now turned nearly towards the earth; but Barnard's conclusion is unlikely to be seriously modified. He fixed the mean diameter of Uranus at 34,900 miles. But this estimate was materially reduced through Dr. See's elimination of irradiative effects by means of daylight measures, executed at Washington in 1901.[1135]

The visual spectrum of this planet was first examined by Father Secchi in 1869, and later, with more advantages for accuracy, by Huggins, Vogel,[1136] and Keeler.[1137] It is a very remarkable one. In lieu of the reflected Fraunhofer lines, imperceptible perhaps through feebleness of light, six broad bands of original absorption appear, one corresponding to the blue-green ray of hydrogen (F), another to the "red-star line" of Jupiter and Saturn, the rest as yet unidentified. The hydrogen band seems much too strong and diffuse to be the mere echo of a solar line, and might accordingly be held to imply the presence of free hydrogen in the Uranian atmosphere. This, however, would be difficult of reconcilement with Keeler's identification of an absorption-group in the yellow with a telluric waterband.

Notwithstanding its high albedo—0·62, according to ZÖllner—proof is wanting that any of the light of Uranus is inherent. Mr. Albert Taylor announced, indeed, in 1889, his detection, with Common's giant reflector, of bright flutings in its spectrum;[1138] but Professor Keeler's examination proved them to be merely contrast effects.[1139] Sir William and Lady Huggins, moreover, obtained about the same time a photograph purely solar in character. The spectrum it represented was crossed by numerous Fraunhofer lines, and by no others. It was, then, presumably composed entirely of reflected light.

Judging from the indications of an almost evanescent spectrum, Neptune, as regards physical condition, is the twin of Uranus, as Saturn of Jupiter. Of the circumstances of his rotation we are as good as completely ignorant. Mr. Maxwell Hall, indeed, noticed at Jamaica, in November and December, 1883, certain rhythmical fluctuations of brightness, suggesting revolution on an axis in slightly less than eight hours;[1140] but Professor Pickering reduces the supposed variability to an amount altogether too small for certain perception, and Dr. G. MÜller denies its existence in toto. It is true their observations were not precisely contemporaneous with those of Mr. Hall[1141] who believes the partial obscurations recorded by himself to have been of a passing kind, and to have suddenly ceased after a fortnight of prevalence. Their less conspicuous renewal was visible to him in November, 1884, confirming a rotation period of 7·92 hours.

It was ascertained at first by indirect means that the orbit of Neptune's satellite is inclined about 20° to his equator. Mr. Marth[1142] having drawn attention to the rapid shifting of its plane of motion, M. Tisserand and Professor Newcomb[1143] independently published the conclusion that such shifting necessarily results from Neptune's ellipsoidal shape. The movement is of the kind exemplified—although with inverted relations—in the precession of the equinoxes. The pole of the satellite, owing to the pull of Neptune's equatorial protuberance, describes a circle round the pole of his equator in a retrograde direction, and in a period of over five hundred years. The amount of compression indicated for the primary body is, at the outside, 1/85; whence it can be inferred that Neptune possesses a lower rotatory velocity than the other giant planets. Direct verification of the trend theoretically inferred for the satellite's movement was obtained by Dr. See in 1899. The Washington 26-inch refractor disclosed to him, under exceptionally favourable conditions, a set of equatorial belts on the disc of Neptune, and they took just the direction prescribed by theory. Their objective reality cannot be doubted, although Barnard was unable, either with the Lick or the Yerkes telescope,[1144] to detect any definite markings on this planet. Its diameter was found by him to be 32,900 miles.

The possibility that Neptune may not be the most remote body circling round the sun has been contemplated ever since he has been known to exist. Within the last few years the position at a given epoch of a planet far beyond his orbital verge has been approximately fixed by two separate investigators.

Professor George Forbes of Edinburgh adopted in 1880 a novel plan of search for unknown members of the solar system, the first idea of which was thrown out by M. Flammarion in November, 1879.[1145] It depends upon the movements of comets. It is well known that those of moderately short periods are, for a reason already explained, connected with the larger planets in such a way that the cometary aphelia fall near some planetary orbit. Jupiter claims a large retinue of such partial dependents, Neptune owns six, and there are two considerable groups, the farthest distances of which from the sun lie respectively near 100 and 300 times that of the earth. At each of these vast intervals, one involving a period of 1,000, the other of 5,000 years, Professor Forbes maintains that an unseen planet circulates. He even computed elements for the nearer of the two, and fixed its place on the celestial sphere;[1146] but the photographic searches made for it by Dr. Roberts at Crowborough and by Mr. Wilson at Daramona proved unavailing. Undeterred by DeichmÜller's discouraging opinion that cometary orbits extending beyond the recognised bounds of the solar system are too imperfectly known to serve as the basis of trustworthy conclusions,[1147] the Edinburgh Professor returned to the attack in 1901.[1148] He now sought to prove that the lost comet of 1556 actually returned in 1844, but with elements so transformed by ultra-Neptunian perturbations as to have escaped immediate identification. If so, the "wanted" planet has just entered the sign Libra, and, being larger than Jupiter, should be possible to find.

Almost simultaneously with Forbes, Professor Todd set about groping for the same object by the help of a totally different set of indications. Adams's approved method commended itself to him; but the hypothetical divagations of Neptune having scarcely yet had time to develop, he was thrown back upon the "residual errors" of Uranus. They gave him a virtually identical situation for the new planet with that derived from the clustering of cometary aphelia.[1149] Yet its assigned distance was little more than half that of the nearer of Professor Forbes's remote pair, and it completed a revolution in 375 instead of 1,000 years. The agreement in them between the positions determined, on separate grounds, for the ultra-Neptunian traveller was merely an odd coincidence; nor can we be certain, until it is seen, that we have really got into touch with it.