Let us compare with Lieutenant Brown's drawing the following description of the corona, in the English Mechanic' for January 27, 1871 (the description was supplied by the special correspondent of that journal): -The corona proper, or glory, or radiated corona, as it is variously called, extended a distance of almost the Moon's diameter from the Moon's limb, but not equally in every direction. It had a greater extension in four directions, at the extremities of two diameters at right angles to each other, so as to give it the shape, roughly speaking, of a square with rounded corners. It was broken in parts, and notably by one decided V-shaped gap. This was observed not only by our party, but at three stations-San Antonio, Xerez, and La Maria Louisa-which form a triangle, each of whose sides is five or six miles in length.'

In Sicily the direct observation of the corona was not so satisfactory as in Spain. Most unfortunately, neither Mr. Lockyer's party nor the party with Professor Roscoe on the slopes of Etna obtained any view at all of the corona. Professor Watson made an excellent drawing of the inner corona, and noticed that the outer was marked by dark rifts. The drawing is highly important, as will presently appear; but it does not afford any information bearing on the nature of the outer and radiated corona.

So much for direct observation. We have clear evidence that at Spanish stations, separated by five or six miles, a certain well-marked feature was seen, and this evidence is not directly contradicted by any obser

vations made at Sicilian stations. When we consider the photographic results, however, all doubts which might be suggested by the absence of any positive evidence from Sicilian stations, in favour of the V-shaped gap, are at once removed.

Fig. 93 is a representation of Mr. Willard's photograph, taken at Cadiz; the limit of the corona, where it extends farthest from the Moon's edge, being due to a stop or diaphragm. Fig. 94 is Mr. Brothers's photograph, taken at Syracuse; the corona is not limited, the image of the Moon having been but threetenths of an inch in diameter on a five-inch plate. It is to be noticed also that Mr. Willard's photograph, being taken with a telescope, required a much longer exposure-in fact, it was exposed for a minute and a half, or nearly the whole time of totality. Mr. Brothers's photograph, on the other hand, was one of a series of five. It was taken during the last eleven seconds of totality, being exposed for eight seconds only. The sky was partially obscured during the earlier part of the total eclipse, but the plates were exposed according to the order planned before the eclipse took place; and with the following results:—

The first and third plates may be regarded as failures

(owing, of course, solely to clouds), the second and

FIG. 93, Mr. Willard's photograph taken near Xeres.
FIG. 94, Mr. Brothers's photograph taken at Syracuse.

D D

fourth, though not judged by Mr. Brothers himself to be so well worth publishing as the fifth, were still well worth studying, and will be presently referred to as affording independent evidence of great importance. The fifth is pictured, as well as a woodcut can picture it, in fig. 93.

It is scarcely necessary for me to indicate the close agreement between these two photographs, or how closely both agree with the recorded results of direct observation. But it must be noted that the woodcuts do not accord nearly so well as the photographs themselves, which, when reduced to the same scale, are absolutely identical as respects the parts of the corona common to both. The figure of the Moon has been impaired in Mr. Brothers's photograph, owing to the swaying of the camera by the wind. The range of prominences and sierra, which had just come into view along the western edge of the Moon, has eaten away' a portion of the Moon's disc on that side; but the corona (whose actinic energy is far smaller than that of the prominences) has not been affected by this circumstance, save in so far as its extension on the western side has been slightly increased by the removal of its base-line eastwards. The notches or gaps round both discs show the places where prominences have by their great actinic energy eaten into the photographic representation of the lunar disc; it will be noticed that their position removes all doubt as to the identity of the coronal rifts. Again, since Mr. Willard's photograph was the result of an exposure lasting during

nearly the whole continuance of totality, it is clear that the rifts were not variable during totality. As Mr. Brothers remarks:- If the great rift had not been persistent, and only became visible during the last twenty seconds of totality at Cadiz, it could not have been shown at all, as the picture would have been "burnt out" at this part, and no dark space would have been shown.'

It is demonstrated then,-first, that a certain wellmarked feature existed in the corona at stations so wide apart as Cadiz and Syracuse; and, secondly, that during the whole continuance of totality at a certain station this feature remained unchanged. The first fact proves incontestably that the V-shaped gap was not caused by any peculiarity in the atmosphere, since we cannot conceive that some peculiarity in the air over Cadiz chanced to have its exact counterpart in the air over Syracuse. The second fact proves as incontestably that the dark rift cannot be due to the passage of the solar rays along lunar irregularities, and thence through any matter (meteoric or otherwise) on our side of the Moon, since any dark rifts so caused would have rapidly changed in shape during the progress of totality.

It cannot be wondered at, therefore, that when Mr. Brothers's photograph was exhibited and described at the meeting of the Royal Astronomical Society, on March 10, 1871, no voice was raised in favour of the atmospheric theory of the corona.