it; in other words, whether the coronal glow shines by its own or by reflected light.

We will briefly review the results obtained by each of these methods. And, first, with regard to the eye-sketches.

A fair number of these, all made, however, at the Spanish end of the shadow-path, have come under our notice, some by artists of pretension, others by draughtsmen of no pretension : and very different are the impressions which these various drawings convey. Clearly, the corona has made greatly varying images upon retinæ of different eyes, or, what is perhaps more likely, the draughtsmen have considerably varied in their powers of conveying their impressions to paper.* Some drawings show a tolerably uniform circle of light, fading equably into the surrounding darkness. Such pictures admit of little inference being drawn from them. Others, on the contrary, exhibit definite points of structure, and where these are apparent in depictions of different observers, we can scarcely doubt the reality of their existence. The tendency of the corona to a roughly quadrangular contour is one feature thus certified; the existence of a definite zone of bright light (which observers of previous eclipses have noted) is another; and it has been suggested that this be named the leucosphere. But perhaps the most remarkable appearance is that of a V-shaped rift in the south-eastern part of the broad coronal haze; and as this is exhibited in several of the drawings, there can be no doubt about its real existence in the corona itself. Another notable feature is presented in three drawings, made by an American observer in Spain, one at the beginning of totality, another in the middle, and a third near the end of totality. In the first of these we see the greatest width of coronal light on the advancing side of the moon; in the second an equal width all around the moon; and in the third an excess on the following side of the moon; the whole series suggesting that the moon acted like a moving shutter intercepting the back-light first

It is probable that much difference may arise from the mere materials of drawing and the fitness of these for the subject. To reproduce a hazy object like the corona, the worst thing to commence with is a sheet of white paper, even if it has a black disc printed upon it to represent the moon. Upon such a tablet great difficulty will arise in working the black sky with the requisite softness around the outer indefinite boundary of the hazy circle, especially if structural details have at the same time to be exhibited. It is in struggling against this difficulty with diverse materials-pencil, chalk, water-colour-that such strangely dissimilar effects are produced. The best basis for reproducing the corona rapidly and effectively would be a sheet of dark grey paper, with a black circle for the moon's disc; and the drawing material should be white chalk, which, upon the grey ground, will produce the desired effects at once with any degree of softness or decision.

on one side, then on the other, from some reflective matter in front of the shutter. This is the prima facie idea conveyed; it is not to be taken as a proffered explanation of the appearance, though it agrees with Oudemans' theory of the coronal beams. Such a transfer of coronal light from one side of the black moon to the other has been depicted before, notably by Professor Plantamour, in his drawings of the eclipse of 1860, July 18, and it led the Astronomer Royal to suggest a cause with which Oudemans' theory is in accordance. Of the photographs secured, three have been exhibited as.

COPY OF MR. BROTHERS' TOTALITY PHOTOGRAPH.*

bearing upon the questions at issue-one taken by Lord Lindsay, in Spain, at the commencement of totality, which is remarkable as exhibiting an excess of corona on the advancing side of the moon, and therefore corroborating the American observer's drawing above noticed; a second obtained by the American party at Xeres, and a third secured by Mr. Brothers at Syracuse. Now, these two last-mentioned photographs are

From a woodcut lent by Messrs. Macmillan. The top represents the north. In viewing this picture regard must be had to the impossibility of reproducing in a woodcut the softness of a photograph.

the most important achievements of the whole eclipse operations. The American one was taken with a six-inch object-glass corrected for actinic rays; the Syracuse one with a photographic copying lens, four inches in diameter, of course mounted equatorially. In the latter there is an extent of corona wider than any other photograph has shown, wider even than that in the corresponding American photograph. No doubt this extension is due to the smallness and consequent brightness of the image, for the exposure was only eight seconds, while the American plate was exposed a minute and a half.* Taken alone, Mr. Brothers' picture is exceedingly valuable: upon one side, for about 120 degrees of the moon's circumference, it shows a spreading of the coronal light to two diameters of the moon; in the southeast part there is the conspicuous rift alluded to as shown by the drawings made in Spain; and there are two other less conspicuous rifts, one about 40 degrees (measured on the moon's circumference) upwards towards the east of the great one, and the other about 60 degrees away to the south. The bright leucosphere is plainly indicated, and a tendency to streaming rays not perfectly radial is decidedly perceptible, even in an enlarged and therefore depreciated copy. But the great value of this picture comes out when it is magnified to equal size and compared with the American one taken at Xeres. Then it is seen that in the main features, especially the conspicuous rifts, the two are almost identical. Slight differences can be made out in the positions of some rather indefinitely-marked portions of the coronal boundary, but on the whole the agreement is wonderfully close. Now, when it is remembered that these pictures were taken at points on the earth's surface 1,100 miles apart, and that one was taken in absolute time 45 minutes after the other, it is perfectly clear that so much of the corona as is common to the two photographs is cosmical and not atmospheric. This is a great point established. And yet those selfdrawn portraits of the corona open out wide fields for speculation. What are we to infer from the wedge-shaped rifts? They seem to deny the possibility of the outer corona being anything like a cosmical cloud near to the moon, on either side, for in that case the moon's motion should have affected them. If they were caused by the shadows of lunar mountains cast upon a sub-lunar mist, in accordance with Oudemans' theory, the motion aforesaid ought to have modified them considerably

In the American picture there is some appearance of the outer coronal light having been cut off as by a diaphragm in the telescope. Structural details are reported to exist in the original negative, which, from causes that will be obvious to a photographer, are not reproduced in an enlarged copy such as that which we have before us.

during the long exposure of the American negative, and the photographed effects of their alteration would have made that picture differ greatly from the one obtained by Mr. Brothers, which had a very brief exposure. And their appearance is incompatible with the supposition of a cosmical cloud near to, though not connected with, the sun; for in that case, why should the rifts maintain a nearly radial position? And if we conceive the corona to be a solar envelope, we cannot regard the gaps as conical openings therein: they must be looked upon as valleys of great extent in the direction of our line of sight. They rather indicate vacant spaces between groups of radiant streamers of luminous or illuminated matter; and this interpretation may well be put upon them by those who have been arguing upon the analogy of the corona to terrestrial aurora and its possible connection therewith-a fascinating subject upon which one would be disposed to dwell if the identity of the coronal with the auroral spectrum lines were less doubtfully established.

corona.

The spectroscopic results from the eclipse are tolerably numerous. The chief point of interest in them is the ample verification of the existence of the green (iron ?) line corresponding to "1474" of Kirchoff's scale, not only in the spectrum of the leucosphere, but in the far outlying regions of the Professor Harkness, at Syracuse, saw the line in all parts as far as 10' from the moon (or sun), and suspected two other green lines less refrangible. Mr. Burton, at Augusta, saw the line also. Professor Winlock, at Xeres, saw it everywhere for 20', or two-thirds the solar diameter, around the sun. Professor Young, also at Xeres, found it half a diameter off. Carpmael, at Estepona, saw three lines, one of which is doubtless the "1474;" we may well infer the same of one of two lines seen by Professor Denza, in Sicily. And we can scarcely doubt that one of those seen by Captain Maclear at San Antonio was also the now famous line; and if so it was, with others (C, D, and F), seen faintly on the moon's disc, thus clearly indicating reflection in our own atmosphere or in some medium between us and the moon. It is thus rendered almost certain that some of the distant coronal haze is reflected light, and this view is strengthened by the facts that Harkness saw a complete hydrogen spectrum when no prominence was near his slit, and that Young saw the c-line far above any possible hydrogen atmosphere.

Enough of observations have probably been secured to fix indubitably the position of the "1474" line. If more information is needed, it would be worth while to try if it cannot be procured from the moon. A spectator upon our satellite at sunrise would see the corona peep above his horizon some

time before the appearance of the actual limb of the sun, and at sunset he would see the corona linger after the sinking of the solar disc. The lunar surface must at sunrise and sunset be illuminated by a coronal twilight, which will be of considerable duration on account of the moon's slow rotation. It is therefore possible that if the faint light seen upon the moon's terminator (the boundary zone of light and darkness) were analysed by the spectroscope it would reveal the coronal lines. At all events, the experiment would be worth trying, and there are abundant opportunities for it. The position of the "1474" line, with respect to lines of known substances, may thus be deliberately determined, and some more reliable evidence obtained to aid a judgment whether it belongs to a new element

-as Professor Young has suggested, some occluded gas, perhaps standing in relation to the magnetic powers of iron of whose spectrum the line is apparently a part-or whether it may be due to iron in any uncommon form or condition.

Reverting to the immediate results of the late eclipse, we remark that a faint continuous spectrum of the corona, without visible dark lines, was noted by several observers, among whom were Lieutenant Brown, Captain Maclear, and Professor Winlock, and that a highly interesting observation, not relating to the corona, however, was made by Professor Young. At the commencement of the totality he saw for an instant the whole of the Fraunhofer lines of the solar spectrum reversed, and the field of his spectroscope filled with bright lines. He must then have caught a glimpse of the stratum of burning elements that lays immediately above the photosphere—an observation made once, and once only, by Lockyer, without an eclipse.

The polariscopic observations confirm those with the spectroscope which indicate that a part of the coronal light is reflected, though they leave open the question whether that reflection occurs in or beyond our atmosphere. Professors Pickering and Langley are reported to have found that a considerable proportion of the light is polarised, and in a radial direction; the first-named observer obtaining the same results with three forms of polariscope. Professor Blaserna, observing in Sicily, asserts that the corona was strongly polarised, the only doubt with him being as to whether the plane was radial to the sun or tangential. Mr. Ranyard, at Villamonda, made three observations, two of which showed what was expected to be observed in the case of radial polarisation. Mr. Pierce, jun., arrived at a similar result, and so did Mr. Ladd. Mr. Samuelson, observing not upon the corona but upon the sky, first, far on one side of the sun, and then far below it, found vertical polarisation.

It would be difficult at the present time to define the precise