"The most imposing as well as complicated of these prominences, which I will call the chandelier, was grand beyond description. It rose up from the limb, appearing like slender tongues of fire, and of a rose colour; its edges purple and transparent, allowing the interior of the prominence to be seen; in fact, I could see distinctly that this protuberance was hollow. Shortly before the end of the totality, I saw escape from the summits of these rosecoloured and transparent sheaves of light, a slight display in the shape of a fan, which gave to the protuberance a real resemblance to a chandelier. Its base, which at the commencement of the totality was noticed very decidedly on the black limb of the moon, became slightly less attached, and the whole took an appearance more ethereal or vapourish; however, I did not lose sight of it for an instant. The jets of light which came from the summits disappeared with the appearance of the first rays of the sun, but it was not so with the protuberance itself, for, an instant before the end of the totality, I saw several small prominences appear lying close to each other on the right of its base, and forming a square, which is the character of toothed prominences; two others of the same height were seen on the left side of its base, when the sun had already appeared, at 2h 55m. It should be mentioned that the images in the field of the telescope were inverted.

"The north horn of the solar crescent touched the last of these prominences, four minutes and forty seconds after the reappearance of the sun, The intense light caused me to abandon this interesting observation, for I was not at the time using a coloured glass; however, I am certain that the chandelier and the little prominences at its base had not disappeared up to that moment.

"Although I am convinced that the protuberances belong to the sun, nevertheless, I ought to remark that, at the last moment, I was surprised to see the direction of the chandelier referred to the centre of the moon, rather than to the centre of the sun. The height of this prominence was estimated about three minutes and a half at the commencement of totality, and four minutes at the end. The second protuberance appeared on the apparent right of this, at a distance of about 35°, being about three minutes and twenty seconds in height, and nearly of the form of the sign of the planet Saturn; this prominence I have called the hook. A third, to the right of the two preceding, and at a distance equal to that of the two others, assumed a form of which it is difficult to give an idea; however, I will call it the tooth. About eleven degrees to the right of the second protuberance I noticed a fourth, small, and in the form of a square; between this and the third there was situated a rose-coloured cloud, the shape of which was elongated and bent, inclined at an angle of 45° towards the left limb of the moon. This cloud was entirely detached, floating on the corona like a red cloud at sunset. Its centre was elevated above the limb of the moon, about one-half the altitude of the other prominences, or about two minutes. A fifth protuberance also appeared, at the beginning, in the south-east, and was of increased size in the middle of the totality.

"I ought to remark, that all the protuberances which I noticed had a tendency in their forms to describe a curve, the concavity of which was turned from the side of the west."

The eclipse was observed by M. Secchi, of the Collegio Romano, at Desierto de las Palmas, near Oropesa, on the eastern coast of Spain, and very nearly on the central line of totality, the estimated

duration of the total darkness at this station being about 3m 26. M. Secchi remarks:

"Shortly before the total disappearance of the sun, I noticed the corona through a lightly coloured glass; I removed the glass as soon as the eclipse became total, and I was astonished at its brilliancy, which was sufficient, even at this time, to dazzle the eye; but its brightness visibly diminished, the limb of the sun being surrounded by a purplish corona, terminating in points of the same colour, which soon disappeared: at this time two magnificent protuberances appeared a little above the spot where the sun's limb disappeared. One was conical, with a point rather slender and curved, having the appearance of a flame somewhat agitated. The other was less elevated, but of greater extent; it occupied an arc of four or five degrees of the limb, the summit terminating like teeth of a very fine saw, the upper outline of which was almost parallel to the limb of the moon. These protuberances. visibly decreased; their height was estimated at 24 and 1 minutes respectively. At the commencement of totality no prominence was visible on the opposite limb of the moon, but about the middle of the eclipse, when the two first had already disappeared, so many luminous points appeared on the other side of the black disk, that I was for a short time embarrassed which to choose for measurement. These brilliant appearances increased in size as fast as the moon glided forward, and I saw with surprise an almost continuous are of purple light instantaneously formed, composed of small protuberances, in that part of the lunar disk where the reappearance of the sun was expected. What surprised me most was a fine red cloud entirely detached from the protuberances, projected on the white light of the corona, and followed by two others of smaller dimensions. I could not refrain from calling the attention of MM. Aguilar and Cepeda, who observed at my side, to this remarkable phenomenon, the existence of which was verified by these observers. Meanwhile, on the side of the lunar disk where the sun was reappearing, the light of the corona gradually increased, and I saw clearly the line on which, in a marked gradation, the white light of the photosphere mingled with the red points of the prominences; the arc, which was tinged with red, extended at that time to at least 60°. Soon after this the protuberances became invisible, but I still saw the corona with the naked eye during 40 seconds after the reappearance of the sun, the solar light shining like an electric lamp, projecting tremulous shadows.

"These observations have convinced me that the protuberances are connected with the sun, and that it is absurd to assert the contrary.”

525. Evidence of a solar atmosphere.—Many of the phenomena attending total solar eclipses afford strong corroboratory evidence of the existence of a solar atmosphere, extending to a vast height above the luminous coating of the sun, the probability of which has been already shown (256).

The corona, or bright ray, or glory, surrounding the dark disk of the moon where it covers the sun, is observed to be concentric with the moon only at the moment when the latter is concentric with the sun. In other positions of the moon's disk, it appears to be concentric with the sun. This would be the effect produced by a solar non-luminous atmosphere faintly reflecting the sun's light.

The corona supplies no exact data by which the height of the solar atmosphere can be ascertained; but it has been the opinion of Sir John Herschel and others, that from the manner in which the diminution of light is manifested on the sun's disk, being by no means sudden on approaching the borders, but extending to some distance within the disk, the height must be not only great in an absolute sense, but must even be a very considerable fraction of the sun's semi-diameter; and this inference is strongly confirmed by the luminous corona surrounding the eclipsed disk.

Although the constitution of the corona is not yet thoroughly understood, the observations made during the eclipses of 1869, 1870, 1871, and 1874, clearly prove that it is a real appendage of the sun, and that for a considerable distance it shines with unborrowed light. Its spectrum, near the chromosphere, shows that it is composed of similar vapours, but the presence of a line identical in position with No. 1474 of Kirchhoff's scale would seem to prove that the vapour of some unknown metal is one of its constituents, of which no indications are found in the spectrum of the chromosphere. From Mr. Stone's spectroscopic examination of the outer corona in the South African eclipse of 1874, which was carried to the extent of rather more than a degree from the sun's centre, we have the strongest evidence of the solar origin and cosmical character of the exterior portions of the corona. This evidence is strengthened by an examination of drawings of the corona made at distances of more than 500 miles, and at intervals of absolute time extending to ten minutes, in which the unchanged character of the principal features is very strongly marked.

526. Probable causes of the red emanations in total solar eclipses. The probable cause of the rose-coloured prominences has been a matter of interesting inquiry for many years, especially since the solar eclipse of 1842. The eclipse of 1851 was very satisfactorily observed, but this question still remained undecided, although it was generally considered that these emanations must be gaseous detached portions of the solar photosphere. The eclipse of 1860 decided that they were of solar origin, which was proved by the photographs taken by Mr. De la Rue and others in Spain at different periods of the totality, on which the progress of the moon over the prominences was so clearly exhibited that their connection with the sun was at once conclusively established.

It was not, however, till the solar eclipse of 1868 occurred, that any certainty of the gaseous character of the prominences was obtained. Between 1860 and 1868, the spectroscope having revealed so much relating to the general physical constitution of the sun, it was considered that the class of observation to be followed in the eclipse of 1868 should be principally confined to

the spectroscopic examination of the solar prominences, with a view of deciding the question as to their solid or gaseous character. The result of this examination proved in a satisfactory manner that these coloured emanations were portions of the chromosphere driven upwards by some violent action in the photosphere, and that their light is emitted principally from the incandescent vapour of hydrogen gas. They appear to have no permanency of existence, for they are continually undergoing changes of form from hour to hour, and as they can now be seen with the spectroscope in ordinary sunlight, these variations of form are even noticed while the prominences are under examination.

II. LUNAR ECLIPSES.

When the moon is in op

527. Cause of lunar eclipses. position, its apparent distance from the plane of the ecliptic or its latitude, varying from o° to upwards of 5°, is at times less than the apparent semi-diameter of the section of the earth's conical shadow, in which case, falling more or less within the shadow, it will be deprived of the sun's light, and will therefore be eclipsed.

The circumstances and conditions attending such a phenomenon depend evidently on the dimensions of the earth's shadow, the magnitude of its section at the moon's distance, and the position of the moon in relation to it.

528. Conditions which determine lunar eclipses. As the earth moves in its orbit round the sun, this conical shadow is therefore constantly projected in a direction contrary to that of the sun. Any body, therefore, which may happen to be in the plane of the ecliptic, or sufficiently near to it, and within this distance of the path of the earth, will be deprived of the sun's light while it is within the limits of the cone. The moon being the only body in the universe which passes within such a distance of the earth, is therefore the only one which can be thus obscured.

The section of the shadow may be regarded as a dark disk, whose apparent semi-diameter varies between 37′ 49′′ and 45′ 42′′, and the true place of whose centre is a point on the ecliptic 180° behind the centre of the sun. A lunar eclipse is produced by the superposition, partial or total, of this disk on that of the moon, and the circumstances and conditions which determine such an eclipse are investigated upon the principles already explained.

By the solar tables, the apparent position of the centre of the sun, from hour to hour, may be ascertained, and the position of the centre of the section of the shadow may thence be inferred. From the lunar tables, the position of the moon's centre being in like manner determined, the distance between the centres of the section

Y

of the shadow and the moon's disk can be ascertained. When this distance is equal to the sum of the apparent semi-diameters of the moon's disk and the section of the shadow, the eclipse will begin; the moment when the distance is least will be the middle of the eclipse, and the line of greatest obscuration; and when the distance between the centres increasing becomes again equal to the sum of the apparent semi-diameters, the eclipse will terminate. The computation of all these conditions, and the time of their occurrence, presents no other difficulty than those of ordinary arithmetical calculation.

The magnitude of the eclipses is measured, like that of the sun, by the difference between the sum of the semi-diameters and the distance between the centres.

The occurrence of a total eclipse, and the moment of its commencement, if it take place, are determined by the distance between the centre of the shadow and that of the moon becoming equal to the difference between the semi-diameter of the shadow and that of the moon. Thus, a total eclipse will take place if the moon's latitude L in opposition be less than

that is, less than the difference between the sum of the horizontal parallaxes and the sum of the semi-diameters; s being the semiangle of the conical shadow, s', the apparent semi-diameter of the section of the shadow at the moon's distance, s, s', the apparent semi-diameters, and h, h', the horizontal parallaxes of the sun and

moon.

Since the sum of the horizontal parallaxes, even when least, is much greater than the sum of the apparent semi-diameters, even when greatest, a total eclipse of the moon is always possible, provided the centre of the moon approaches near enough to the centre of the shadow, and for the same reason an annular lunar eclipse is impossible.

529. Lunar ecliptic limits. That a lunar eclipse may take place, it is necessary that the moon, when in opposition, should approach the ecliptic within a distance less than the sum of the apparent semi-diameters of the moon and the section of the shadow. Let its latitude in opposition be L', the limiting value of this will be

L'=h+h'+s'— 8.

If the latitude of the moon be less than this (which is the sum of the semi-diameters of the moon and shadow) an eclipse must take place.

But, as in the case of solar eclipses, the quantities composing