The twelvefold division of the zodiac was evidently suggested by the occurrence of twelve full moons in successive parts of it in the course of each year. This approximate relation was first systematically developed by the early inhabitants of Mesopotamia, and formed the starting-point for all their divisions of time. As the year separated, as it were of itself, into twelve months, so the day was divided into twelve “double hours,” and the great cosmical period of 43,200 years into twelve “sars.” Each sar, month, and hour was represented at once visibly and symbolically by a twelfth part of the “furrow” drawn by the solar Bull across the heavens. The idea of tracing the sun’s path among the stars was, when it occurred to Chaldaean astronomers, an original and, relatively to their means, a recondite one. We owe to its realization by them the constitution and nomenclature of the twelve signs of the zodiac. Assyrian cylinders and inscriptions indicate for the familiar series of our text-books an antiquity of some four thousand years. Ages before Asurbanipal reigned at Nineveh the eighth month (Marchesvan) was known as “the month of the star of the Scorpion,” the tenth (Tebet) belonged to the “star of the Goat,” the twelfth (Adar) to the “star of the Fish of Hea.” The motive underlying the choice of symbols is in a few cases obvious, but in most remains conjectural. The attributes of the deities appointed to preside over the months and signs were to some extent influential. Two of them, indeed, took direct possession of their respective portions of the sky. The zodiacal Virgo is held to represent the Assyrian Venus, Ishtar, the ruling divinity of the sixth month, and Sagittarius the archer-god Nergal, to whom the ninth month was dedicated. But no uniform system of selection was pursued; or rather perhaps the results of several, adopted at various epochs, and under the influence of varying currents of ideas, became amalgamated in the final series.

This, there is reason to believe, was the upshot of a prehistoric reform. So far as positive records go, Aries was always the first sign. But the arrangement is, on the face of it, a comparatively modern one. None of the brighter stars of the constellation could be said even roughly to mark the equinox much before 1800 b.c.; during a long stretch of previous time the leading position belonged to the stars of Taurus. Numerous indications accordingly point to a corresponding primitive zodiac. Setting aside as doubtful evidence derived from interpretations of cuneiform inscriptions, we meet, in connexion with Mithraic and Mylittic legends, reminiscences of a zodiac and religious calendar in which the Bull led the way. Virgil’s

Candidus auratis aperit cum cornibus annum Taurus perpetuates the tradition. And we shall see presently that the Pleiades, not only were originally, but continued to be until well within historical memory, the first asterism of the lunar zodiac.

In the Chaldaean signs fragments of several distinct strata of thought appear to be embedded. From one point of view, they shadow out the great epic of the destinies of the human race; again, the universal solar myth claims a share in them; hoary traditions were brought into ex. post facto connexion with them; or they served to commemorate simple meteorological and astronomical facts.

The first Babylonian month Nisan, dedicated to Anu and Bel, was that of “sacrifice”; and its association with the Ram as the chief primitive object of sacrifice is thus intelligible. According to an alternative explanation, however, the heavenly Ram, placed as leader in front of the flock of the stars, merely embodied a spontaneous figure of the popular imagination. An antique persuasion, that the grand cycle of creation opened under the first sign, has been transmitted to modern cognizance by Dante (Inf, i. 38). The human race, on the other hand, was supposed to have come into being under Taurus. The solar interpretation of the sign goes back to the far-off time when the year began with Taurus, and the sun was conceived of as a bull entering upon the great furrow of heaven as he ploughed his way among the stars. In the third month and sign the building of the first city and the fratricidal brothers—the Romulus and Remus of Roman legend— were brought to mind. The appropriate symbol was at first indifferently a pile of bricks or two male children, always on early monuments placed feet to feet. The retrograde movement of a crab typified, by an easy association of ideas, the retreat of the sun from his farthest northern excursion, and Cancer was constituted the sign of the summer solstice. The Lion, as the symbol of fire, represented the culmination of the solar heat. In the sixth month, the descent of Ishtar to Hades in search of her lost husband Tammuz was celebrated, and the sign of the Virgin had thus a purely mythological signification.

The history of the seventh sign is somewhat complicated. The earlier Greek writers,—Eudoxus, Eratosthenes, Hipparchus,—knew of only eleven zodiacal symbols, but made one do double duty, extending the Scorpion across the seventh and eighth divisions. The Balance, obviously indicating the equality of day and night, is first mentioned as the sign of the autumnal equinox by Geminus and Varro, and obtained, through Sosigenes of Alexandria, official recognition in the Julian calendar. Nevertheless, Virgil (Georg., i. 32) regarded the space it presided over as so much waste land, provisionally occupied by the “Claws” of the Scorpion, but readily available for the apotheosis of Augustus. Libra was not of Greek invention. Ptolemy, who himself chiefly used the “Claws” (X ηλαί ), speaks of it as a distinctively Chaldaean sign; and it occurs as an extra-zodiacal asterism in the Chinese sphere. An ancient Chinese law, moreover, prescribed the regularization of weights and measures at the spring equinox. No representation of the seventh sign has yet been discovered on any Euphratean monument; but it is noticeable that the eighth is frequently doubled, and it is difficult to avoid seeing in the pair of zodiacal scorpions carved on Assyrian cylinders the prototype of the Greek scorpion and claws. Both Libra and the sign it eventually superseded thus owned a Chaldaean birthplace. The struggle of rival systems of nomenclature, from which our zodiacal series resulted, is plainly visible in their alternations; and the claims of the competing signs were long sought to be conciliated by representing the Balance as held between the claws of the Scorpion.

The definitive decline of the sun’s power after the autumnal equinox was typified by placing a Scorpion as the symbol of darkness in the eighth sign. Sagittarius, figured later as a Centaur, stood for the Babylonian Mars. Capricornus, the sign of the winter solstice, is plausibly connected with the caprine nurse of the young solar god in Oriental legends, of which that of Zeus and Amalthea is a variant. The fish-tailed Goat of the zodiac presents a close analogy with the Mexican calendar sign Cipactli, a kind of marine monster resembling a narwhal. Aquarius is a still more exclusively meteorological sign than Leo. The eleventh month was known in Euphratean regions as that of “want and rain.” The deluge was traditionally associated with it. It was represented in zodiacal symbolism by the god Ramman, crowned with a tiara and pouring water from a vase, or more generally by the vase and water without the god. The resumption of agricultural labours after the deluge was commemorated in the twelfth month, and a mystical association of the fishes, which were its sign, with the life after death is evident in a monument of Assyrian origin described by M. Clermont-Ganneau, showing a corpse guarded by a pair of fish-gods. The doubling of the sign of Pisces still recalls, according to Mr Sayce, the arrangement of the Babylonian calendar, in which a year of 360 days was supplemented once in six years by a thirteenth month, a second Adar. To the double month corresponded the double sign of the “Fishes of Hea.”

The cyclical meaning of the succession of zodiacal signs, though now obscured by interpolations and substitutions, was probably once clear and entire. It is curiously reflected in the adventures of the Babylonian Hercules, the solar hero Izdubar. They were recorded in the comparatively late surviving version of the 7th century b.c., on twelve tablets, with an obvious design of correlation with the twelve divisions of the sun’s annual course. Izdubar’s conquest of the winged bull Heabani was placed under Taurus; his slaying of the tyrant Houmbaba (the prototype of Geryon) in the fifth month typified the victory of light over darkness, represented in plastic art by the group of a lion killing a bull, which is the form ordinarily given to the sign Leo on Ninevite cylinders. The wooing of Ishtar by the hero of the epic falls under Virgo, and his encounter with two scorpion men, guardians of the rising and the setting sun, under Scorpio. The eleventh tablet narrates the deluge; the twelfth associates the apotheosis of Heabani (the Babylonian Chiron) with the zodiacal emblems of the resurrection.

In the formation of the constellations of the zodiac very little regard was paid to stellar configurations. The Chaldaeans chose three stars in each sign to be the “councillor gods” of the planets. These were called by the Greeks “decans,” because ten degrees of the ecliptic and ten days of the year were presided over by each. The college of the decans was conceived as moving, by their annual risings and settings, in an “eternal circuit” between the infernal and supernal regions. Our modern asterisms first appear in the Phaenomena of Eudoxus about 370 b.c. But Eudoxus, there is reason to believe, consulted, not the heavens, but a celestial globe of an anterior epoch, on which the stars and the signs were forced into unnatural agreement. The representation thus handed down to us (in the verses of Aratus) has been thought to tally best with the state of the sky about 2000 b.c.; and the mention of a pole-star, for which Eudoxus was rebuked by Hipparchus, seems, as Mr W. T. Lynn has pointed out, to refer to the time when α Draconis stood near the pole. The data afforded by Eudoxus, however, are far too vague to serve as the basis of any chronological conclusion.

The Egyptians adopted from the Greeks, with considerable modifications of its attendant symbolism, the twelvefold division of the zodiac. Aries became the Fleece; two Sprouting Plants, typifying equality or resemblance, stood for Gemini; Cancer was re-named Scarabaeus; Leo was converted, from the axe-like configuration of its chief stars, into the Knife; Libra into the Mountain of the Sun, a reminiscence, apparently, of the Euphratean association of the seventh month with a “holy mound,” designating the Biblical tower of Babel. A Serpent was the Egyptian equivalent of Scorpio; the Arrow only of Sagittarius was retained; Capricornus became “Life,” or a Mirror as an image of life; Aquarius survived as Water; Taurus, Virgo, and Pisces remained unchanged. The motive of some of the substitutions was to avoid the confusion which must have ensued from the duplication of previously existing native asterisms; thus, the Egyptian and Greek Lions were composed of totally different stars. Abstractions in other cases replaced concrete objects, with the general result of effacing the distinctive character of the Greek zodiac as a “circle of living things.”

Early Zoroastrian writings, though impregnated with star-worship, show no traces of an attempt to organize the heavenly array. In the Bundehish, however (9th century), the twelve “Akhtârs,” designated by the same names as our signs, lead the army of Ormuzd, while the seven “Awakhtârs” or planets (including a meteor and a comet) fight for Ahriman. The knowledge of the solar zodiac thus turned to account for dualistic purposes was undoubtedly derived from the Greeks. By them, too, it was introduced into Hindustan. Aryabhata, about the beginning of our era, reckoned by the same signs as Hipparchus. They were transmitted from India by Buddhist missionaries to China, but remained in abeyance until the Jesuit reform of Chinese astronomy in the 17th century.

The native zodiacal system was of unexampled complexity. Besides divisions into twenty-eight and twenty-four parts, it included two distinct duodenary series. The tse or “stations” were referred by Biot to the date 1111 b.c. Measured from the winter solstice of that epoch, they corresponded, in conformity with the Chinese method of observation by intervals of what we now call right ascension, to equal portions of the celestial equator. Projected upon the ecliptic, these were, of course, considerably unequal, and the tse accordingly differed essentially from the Chaldaean and Greek signs. Their use was chiefly astrological, and their highly figurative names—“Great Splendour,” “Immense Void,” “Fire of the Phoenix,” &c. —had reference to no particular stars. They became virtually merged in the European series, stamped with official recognition upwards of two centuries ago. The twenty-four tsieki or demi- tse were probably invented to mark the course of weather changes throughout the year. Their appellations are purely meteorological.

The characteristic Chinese mode of dividing the “yellow road” of the sun was, however, by the twelve “cyclical animals,”—Rat, Ox, Tiger, Hare, Dragon or Crocodile, Serpent, Horse, Sheep, Monkey, Hen, Dog, Pig. The opening sign corresponds to our Aquarius, and it is remarkable that the rat is, in the far East, frequently used as an ideograph for “water.” But here the agreement ceases. For the Chinese series has the strange peculiarity of proceeding in a retrograde direction or against the course of the sun. Thus, the second sign (of the Ox) occupies the position of Capricorn, the third that of Sagittarius, and so on. The explanation of this seeming anomaly is to be found in the primitive destination of the “animals” to the purposes of an “horary zodiac.” Their succession, established to mark the hours of day and night, was not unnaturally associated with the diurnal revolution of the sphere from east to west. They are unquestionably of native origin. Tradition ascribes their invention to Tajao, minister of the emperor Hwang-te, who reigned c. 2697 b.c., and it can scarcely be placed later than the 7th century b.c.Z

The Chinese circle of the “animals” obtained early a wide diffusion. It was adopted by Tartars, Turks, and Mongols, in Tibet and Tong-king, Japan and Corea. It is denominated by Humboldt the “zodiac of hunters and shepherds,” and he adds that the presence in it of a tiger gives it an exclusively Asiatic character. It appears never to have been designed for astronomical employment. From the first it served to characterize the divisions of time. The nomenclature not only of the hours of the day and of their minutest intervals was supplied by it, but of the months of the year, of the years in the Oriental sixty-year cycle, and of the days in the “little cycle” of twelve days. Nor has it yet fallen into desuetude. Years “of the Rat,” “of the Tiger,” “of the Pig,” still figure in the almanacs of Central Asia, Cochin China, and Japan.

A large detachment of the “cyclical animals” even found its way to the New World. Seven of the twenty days constituting the Aztec month bore names evidently borrowed from those of the Chinese horary signs. The Hare (or Rabbit), Monkey, Dog, and Serpent reappeared without change; for the Tiger, Crocodile, and Hen, unknown in America, the Ocelot, Lizard, and Eagle were substituted as analogous. The Aztec calendar dated from the 7th century; but the zodiacal tradition embodied by it was doubtless much more ancient. Of the zodiac in its true sense of a partitioned belt of the sphere there was no aboriginal knowledge on the American continent. Mexican acquaintance with the signs related only to their secondary function as dies (so to speak) with which to stamp recurring intervals of time.

The synodical revolution of the moon laid down the lines of the solar, its sidereal revolution those of the lunar zodiac. The first was a circlet of “full moons”; the second marked the diurnal stages of the lunar progress round the sky, from and back again to any selected star. The moon was the earliest “measurer” both of time and space; but its services can scarcely have been rendered available until stellar “milestones” were established at suitable points along its path. Such were the Hindu nakshatras, a word originally signifying stars in general, but appropriated to designate certain small stellar groups marking the divisions of the lunar track. They exhibit in an exaggerated form the irregularities of distribution visible in our zodiacal constellations, and present the further anomaly of being frequently reckoned as twenty-eight in number, while the ecliptical arcs they characterize are invariably twenty-seven. Now, since the moon revolves round the earth in 27⅓ days, hesitation between the two full numbers might easily arise; yet the real explanation of the difficulty appears to be different. The superfluous asterism, named Abhijit, included the bright star α Lyrae, under whose influence the gods had vanquished the Asuras. Its invocation with the other nakshatras, remoteness from the ecliptic notwithstanding, was thus due (according to Prof. Max Müller’s plausible conjecture) to its being regarded as of especially good omen. Acquaintance with foreign systems of twenty-eight lunar divisions tended doubtless to fix its position, which remained, nevertheless, always equivocal. Alternately admitted into or rejected from the series, it was finally, some six or seven centuries ago, eliminated by the effects of precession in reversing the order of culmination of its limiting stars.

The notion of a twenty-seven-fold division of the zodiac was deeply rooted in Hindu tradition. The number and the name were in early times almost synonymous. Thus, a nakshatra-mālā denoted a necklace of twenty-seven pearls; and the fundamental equality of the parts was figured in an ancient legend, by the compulsion laid upon King Soma (the Moon) to share his time impartially between all his wives, the twenty-seven daughters of Pra-jápati. Everything points to a native origin for the system of nakshatras. Some were named after exclusively Vedic deities; they formed the basis of the sacrificial calendar of the Brahmans; the old Indian names of the months were derived from them; their existence was presupposed in the entire structure of Hindu ritual and science. They do not, however, obtain full recognition in Sanskrit literature until the Brāhmana period (7th or 8th century b.c.). The Rig-Veda contains only one allusion to them, where it is said that “Soma is placed in the lap of the nakshatras ”; and this is in a part including later interpolations.

Positive proof of the high antiquity of the Hindu lunar zodiac is nevertheless afforded by the undoubted fact that the primitive series opened with Krittikā (the Pleiades) as the sign of the vernal equinox. The arrangement would have been correct about 2300 B.c.; it would scarcely have been possible after 1800 B.c. We find nowhere else a well-authenticated zodiacal sequence corresponding to so early a date. The reform by which Krittikā, now relegated to the third place, was superseded as the head of the series by “Açvini” was accomplished under Greek influence somewhere near the beginning of our era. For purposes of ritual, however, the Pleiades, with Agni or “Fire” as their presiding deity, continued to be the first sign. Hindu astronomy received its first definite organization in the 6th century, with results embodied in the Sūrya-Siddhānta. Here the “signs’’and the “constellations” of the lunar zodiac form two essentially distinct systems. The ecliptic is divided into twenty-seven equal parts, called bhogas or arcs, of 800' each. But the nakshatras are twenty-eight, and are represented by as many “junction stars” (yogātāra), carefully determined by their spherical coordinates. The successive entries of the moon and planets into the nakshatras (the ascertainment of which was...

Among the Moslems, to whom it is important on ritual grounds to determine accurately the moment of daybreak, at which during Ramadan the daily fast begins, the morning zodiacal light appears to have been observed from an early period, and is known as the “false dawn” or the “wolf’s tail” (Redhouse, in Journ. R. A. S., July 1878). But in Christian Europe it seems to have been first observed by Kepler, who described its appearance with considerable accuracy and came to the conclusion that it was the atmosphere of the sun. Descartes wrote about it in 1630 and Childrey in 1659; but the attention of astronomers was first prominently called to it by Dominic Cassini, who first saw it on 18th March 1683. It is to him that it owes the name which it now bears. He explained it by supposing the existence of a flat, luminous ring encircling the sun, nearly in the plane of his equator, and accounted for its disappearance on the same principle as that which accounts for the vanishing of Saturn’s ring. Mairan (1731), like Kepler, ascribed the light to the sun’s atmosphere; and this explanation was generally accepted, till Laplace showed that it was untenable, since no real solar atmosphere could extend to anything like the distance from the sun which is reached by the zodiacal light. He further showed that, even if the solar atmosphere did extend far enough, it would not have the lenticular appearance ascribed by observers to the zodiacal light, since the polar axis would be at least two-thirds of the equatorial axis. Since then many observers have made a study of the subject, amongst whom may be mentioned Jones, Piazzi Smyth, Jacob, Brorsen, Schmidt, Backhouse, Liais, and Wright.

Extent.— The way in which the light fades off gradually towards the boundaries makes it extremely difficult to determine accurately the true position of the light or its extent. Various observations show that at times the base, at an elongation of about 20°, may have a width of from 25° to 30°, while at an elongation of 60° the breadth is frequently as much as 20°, but usually much less. The distance of the vertex from the sun frequently exceeds 90°, and Mr Liais and others have recorded cases when the light has been traced completely round from the western to the eastern horizon. This is very uncommon; but it is not at all rare to find the light stretching nearly to the meridian three hours after sunset, and several observers have recorded the existence of a bright patch of light almost opposite to the position of the sun. This is known as the Gegenschein, and though it has been seen comparatively seldom its existence must be accepted as proved; for its position has been determined by actual measurement by several astronomers, and their results agree with quite as great closeness as can be expected in the determination of the position of such an object. A lengthened series of observations was made on the zodiacal light by the Rev. G. Jones, chaplain of the United States steam frigate “Mississippi,” in the China and Japan Seas in 1855. He charted the apparent position of the cone of light on a large number of nights and mornings, and came to the somewhat startling conclusion that his observations could be explained only by supposing the existence of a nebulous ring round the earth within the orbit of the moon. He recorded that twice near 23° 28' N. lat., with the sun at the opposite solstice, he had seen “the extraordinary spectacle of the zodiacal light, simultaneously at both the east and west horizons from 11 to 1 o’clock for several nights in succession.” On reading this statement, Baron Humboldt communicated to the Monatsberichte d. kön. preuss. Akad. d. Wiss. some unpublished observations of his own on a similar phenomenon. These were to the effect that on the 17th and 18th of March the light was very bright in the west, and “we constantly perceived in the east (and this is beyond doubt a very striking phenomenon) a whitish light which was also of a pyramidal form. The latter augmented the brightness of the sky in a very striking manner.” The light in the west was so conspicuous that “even the sailors were delighted with this double light.” He noted, too, what is a very important fact, that the two lights set at the same time; and there is much to be said in defence of his view that the eastern light was the reflexion of the true zodiacal light in the west, just as the eastern sunset glows are the reflexion of those in the west. The chief obstacle in the way of accepting this view is that it is difficult to believe that a light so faint as the zodiacal light could have a reflexion bright enough to be seen, and even to be seen distinctly. On the other side must be set the circumstance that the simultaneous glow at both horizons seems never to have been observed from great altitudes. This is easily understood if the second glow is a mere reflexion; whereas, if it comes from another source of light, it ought to be more conspicuous at high than at low levels.

Position.— The exact position of the axis of the zodiacal light relatively to the ecliptic has not yet been satisfactorily fixed. The extreme haziness of outline and the excessive faintness of the light near the vertex make it quite impossible to use a telescope for measuring its limits. Most observers have tried to fix its position by tracing its outline on a star chart, while Prof. C. Piazzi Smyth employed two sights mounted equatorially. But even by these means no great accuracy can be attained, for the limits of the light can be traced only when the eye is quite unfatigued and when the light is looked at with averted vision. The difficulty experienced is well illustrated by the wide divergencies between the results of different observers—divergencies not only in the extent of the light, which would be quite natural, but even in the direction of the axis of the cone. Thus, Captain Jacob, who observed in Madras in 1856- 57-58, found that his evening observations placed the vertex of the western cone in from 2° S. lat. to 6° N. lat., with an average of about 3° N. lat. His morning observations were less numerous and gave positions for the vertex of the eastern cone varying from ½ o N. to 8° S., with an average of about 2° S. Mr Backhouse from 418 observations found a mean deviation from the ecliptic of 2°∙06, while Captain Tupman, observing in the Mediterranean in August and September, found an inclination of no less than 20°. He also found that the plane of the light did not pass through the sun. His observations, however, differ somewhat widely from those of most other observers. Mr A. Searle has made a very careful study of all the best published records. In his first paper more than 650 observations by Jones, Heis, Schmidt, and others are discussed, and he concludes that apparent changes in latitude are mainly, if not entirely, produced by the effect of atmospheric absorption, which affects the lower boundary more than the upper and to an extent depending on the inclination of the axis of the light to the horizon. In a second paper he discusses Jones’s observations alone and comes to the conclusion that, after making allowance for the effects of absorption, there is some evidence that the zodiacal light, as seen during the second half of the 19th century, has had a more northern latitude near 180° long. than near 0° long. It seems somewhat doubtful, however, whether the observations hitherto made are sufficiently accurate to justify this conclusion, and all that can be confidently asserted is that the observations made so far point to such a difference as being probable. Similar variations in the greatest elongation, the breadth, and the brightness have been asserted to exist. Thus Mr J. F. Julius Schmidt, as the result of several years’ observations, finds a variation in the inclination of the axis of the cone of light to the ecliptic from 4° 18' at the end of December to 0° towards the end of March, a similar variation in the greatest elongation from the sun from 120° in January to 70° in April, and a further similar variation in breadth at corresponding elongations, with a minimum for all three about 30th March. It may be pointed out that observations regarding the extent of the visible light are absolutely valueless unless accompanied by careful records of the clearness of the sky as evidenced by the visibility of faint objects, for all observers of the zodiacal light must have noticed that a thin film of mist, which hardly affects the colour of the sky or the visibility of the brighter stars, may almost completely obliterate the zodiacal light. No accurate photometric observations seem to have been made on the strength of the light and observers have differed widely in their estimates. Some assert that it is much brighter than the Milky Way, while others—and especially those who have observed in the tropics— say that it is seldom so bright as the brighter parts of the Milky Way. It is certainly often brighter than most of the Milky Way. Mr Searle deduces from Sir W. Herschel’s and Celoria’s observations that the Milky Way is about 2 magnitudes brighter than the mean brightness of the sky, and says that on this estimate the brighter parts of the zodiacal light would be commonly 3 or 4 magnitudes brighter than the surrounding sky. This is almost certainly an over-estimate. There seems to be very little doubt that the brightness undergoes periodic fluctuations, but no estimate can as yet be made of the length of the period. It is probable that the brightness has been below the average for the last few years, but is now increasing. Most observers have also reported rapid changes of brightness, or undulations, such as are seen in the aurora and in the tails of some comets; but, as was pointed out by Others in 1833, these undulations must be produced in the earth’s atmosphere. In this connexion it may be mentioned that in observing the spectrum rapid flickerings, like waves moving along the spectrum, are often very marked.

The Moon’s Zodiacal Light.— Several observers have recorded observations which appeared to show that the moon produced an appearance very similar to that of the zodiacal light. Piazzi Smyth, however, when observing on the Peak of Teneriffe, saw this appearance and showed by actual measurement that the glow seen before moonrise does not lie near the ecliptic, but is nearly vertical, and is due simply to refraction in the earth’s atmosphere. This explanation will hardly account for an interesting observation made by Mr L. Trouvelot, which if repeated would require to be very carefully investigated. On a night when the zodiacal light was very bright and there were magnetic disturbances followed by an auroral display, but when no aurora was actually visible, he saw a conical light rising obliquely from the top of the roof of a building behind which the moon, then about 15° or 20° above the horizon, was concealed. The axis of the light coincided nearly with the ecliptic and the light could be traced on both sides of the moon, when the moon itself was concealed. The whole of the circumstances led Mr Trouvelot to conclude that this light and the zodiacal light were phenomena of the same order, while this and other observations, he considered, rendered it probable that there was some connexion between the zodiacal light and auroras.

Physical Constitution.— As has already been pointed out, it is impossible to see the zodiacal light through a telescope, and this, taken along with the extreme faintness of the light, renders it exceedingly difficult to examine it satisfactorily with either the polariscope or the spectroscope. Many attempts have been made to determine whether or not the light was to any extent polarized, but with questionable results until Prof. A. W. Wright attacked the problem, using a polariscope specially designed for studying very faint lights. With this he was enabled to determine with certainty that the light was partially polarized in a plane passing through the sun, and that the amount of polarization was most probably as much as 15 per cent., but less than 20 per cent. Many attempts have been made to observe the spectrum. In 1867 Ångström, observing at Upsala in March, obtained the bright aurora line (W. L. 5567), and concluded that in the zodiacal light there was the same material as is found in the aurora and in the solar corona, and probably through all space. Upsala, however, is a place where the aurora spectrum can often be observed in the sky even when no aurora is visible, and it has generally been believed that what Ångström really saw was an auroral and not a zodiacal light spectrum. Señor A. T. Acrimiz, observing at Cadiz, obtained a continuous but faint spectrum with two bright lines—a yellowish line, probably an aurora line, and a line in the blue, more refrangible than F, which he could not identify, but which in all probability was another auroral line. The fact that he saw this spectrum with a five-prism spectroscope attached to an equatorial seems conclusive evidence that it was not the zodiacal light spectrum that he was observing. The most satisfactory observations hitherto published seem to be those of Prof: Piazzi Smyth and Prof. A. W. Wright. Both used spectroscopes specially designed for the examination of faint lights, and their results agree completely with each other. Prof. Smyth made his observations at Palermo and found a faint continuous spectrum extending from about W. L. 5550 to W. L. 5000 (British inches scale), with a maximum brightness at about W. L. 5350. In fact, the light was almost exactly similar to that of faint diffused sunlight, such as is got in the last traces of twilight. Prof. Wright’s conclusion was that the spectrum differs from that of sunlight only in intensity. Some recent unpublished observations made in the tropics indicate that, while the spectrum is usually that described by Prof. Smyth, there are times when a bright line is seen. Too much stress, however, must not be laid on this, as the observations are by no means conclusive, and the apparent line may in reality be only a part of the continuous spectrum which is brighter than usual.

The discussion of the real cause of the zodiacal light is rendered very difficult by the want of agreement in the observations that have been made upon it, and by the existence of a small number of apparently trustworthy observations of a very abnormal extension of the light, as detailed above; but certain conclusions may be safely arrived at. The theory that it is due to a ring of small bodies surrounding the earth seems to be entirely negatived, as pointed out by Proctor and others, by several of its features. The best observations leave no room for the parallactic displacement which would be observed if there were such a ring round the earth, and the absence of a large part of the luminous circle on ordinary occasions would be inexplicable on any such hypothesis. There can therefore be very little doubt that we must look for the cause of the light to the existence of a mass of small bodies moving in orbits round the sun, and that, as shown by the polarization and the spectrum, the light is chiefly, if not entirely, reflected sunlight. To account for all the observations, Proctor has shown that the bodies must travel in orbits of considerable eccentricity, carrying them far beyond the limits of what we may term the zodiacal disk. Under these conditions the constitution of the disk would become variable within exceedingly wide limits, and all the recorded variations might be fairly explained. The appearance of a complete arch, as seen by Liais and others, would indicate an extension far beyond the earth’s orbit; but, at present at least, this must be looked on as very exceptional. The spectroscopic evidence, so far as it goes, confirms this theory, since it indicates that the light is mainly reflected sunlight; and, even if further observations should confirm the opinion that bright lines are sometimes present, this need in no way invalidate the conclusions that have been arrived at. Indeed, it seems highly probable—especially if Mr Lockyer’s views regarding meteorites are confirmed—that bright lines should be seen in such a body of meteorites. Olbers long ago suggested that the corona was the brightest part of the zodiacal light, and there is nothing in recent observations to contradict this view, for all observations go to show that the outer corona has no definite boundary, but shades off imperceptibly and becomes invisible at a great distance from the sun. Dr Huggins, while holding that the corona is most probably due to the ceaseless outflow of extremely minute particles from the sun, thinks it not improbable that the zodiacal light may be in some way connected with this outflow. Dr Siemens, when discussing his theory of the conservation of solar energy, sought for an explanation of the zodiacal light in the dust which he supposed to be ejected from equatorial regions, rendered luminous partly by reflected sunlight, partly by phosphorescence, and partly by electrical action.

With the increasing number of observatories at high altitudes it may fairly be hoped that before long astronomers will be put in possession of such definite measurements as will enable some at least of the points still under discussion to be finally settled, and that far more accurate observations will soon be available on which to construct a satisfactory theory.

Bibliography.— Childrey, Natural History of England (1659) and Britannia Baconica, p. 183 (1661); Cassini (D.), Nouv. Phénom. d'une lumière céleste [zodiacale] (1683) and Découverte de la lumière céleste qui paroist dans le zodiaque (1685); Hooke (R.), Explication of a Glade of Light, &c. (1685); Mairan, Observations de la lumière zodiacale ; Euler (L.), Sur la cause de la lumière zodiacale (1746); Mairan, Sur la cause de la lumière zodiacale (1747); Wolf (R.), Beobachtungen des Zodiacallichtes...