like 2 and 5 in our figure. Sixteen years or so earlier or later the November system would present a similar appearance, only very much fainter, on account of greatly increased distance, during a total eclipse occurring on or about November 13. At no other time in the year except November 13 and May 10, or about these dates, could the November system present such an appearance. But a system travelling close to the sun, and not far from the plane near which all the planets travel, would present at all times nearly the appearance of a pair of rays like 2 and 5 of our figure. On this account, therefore, as well as on account of the greater brightness with which such meteor systems would be illuminated, we must prefer the theory that the systems to which the coronal rays are due travel near to the sun.

Yet, even as thus presented, the meteor theory alone seems inadequate to explain the coronal streamers. There is an enormous mass of evidence showing that meteor systems are most richly strewn throughout a region around the sun extending nearly to the distance of the planet Mercury; but there is also abundant reason for believing that these multitudinous systems would present an appearance very different from that depicted in Professor Abbe's view of the coronal streamers. We want something quite distinct from the theory of a mere aggregation of meteors to account for these rays, whether pointed or fan-shaped, extending directly from the sun. The aggregation of meteors might present the appearance of a luminous cloud around the place of the eclipsed sun. This cloud might be to some degree radiated, because each meteor system would have a course carrying it either directly athwart the sun's place on the sky, or nearly so. But there would be nothing like those sharply-defined streamers extending separately from the sun to distances of ten or twelve sun-breadths. Sir George Airy, describing the appearance of the corona during the eclipse of 1851, pictures just such a cloud as we should expect to result from the aggregation of meteors.

'Its colour,' he said, 'was white, or resembling that of Venus; there was no flickering or unsteadiness; it was not separated from the moon, nor had it any annular structure: it looked like a radiated luminous cloud behind the moon. The long streamers manifestly require a different explanation.

I cannot but think that the true explanation of these streamers, whatever it may be (I am not in the least prepared to say what it is), will be found whensoever astronomers have found an explanation of comets' tails. These singular appendages, like the streamers seen by Professor Abbe, extend directly from the sun, as if he exerted some repellent action on the matter forming the heads of comets. Indeed, Sir John Herschel did not hesitate to say that the existence of such a repulsive force was, to all intents and purposes, demonstrated by the phenomena of comets' tails. Now we know that meteors and comets are in some way associated, though the actual nature of the connection between them is not clear. It is certain that the November meteors, the August meteors, and other such systems, follow in the track of known comets. We know that when, in 1862, the earth passed through the region of space along which Biela's comet had recently travelled, there was a display of thousands of meteors, all radiating from just that part of the heavens from which bodies travelling parallel to the orbit of Biela's comet would have seemed to radiate. It follows from this association between comets and meteors, and from the fact that probably thousands of meteoric and cometic systems travel close to the sun, that in all probability there must exist generally, if not always, in the sun's neighbourhood, enormous quantities of the substance whence comets' tails are formed by the sun's repellent action. This being so, we should expect to find generally, if not always, long streams of matter extending from the sun's immediate neighbourhood, in the same way that comets' tails extend from comets' heads. Whether the repulsive force is electrical, magnetic, or otherwise, does not at present concern

us, or rather it does concern us, but at present we are quite unable to answer the question. All that we know certainly is that, in the first place, the sun does in some way cause streams of luminous matter to appear beyond the heads of comets, in a direction opposite to his own, and to enormous distances; and, in the second place, that the matter forming comets' heads is probably present at all times, in large quantities, in the sun's immediate neighbourhood. We can hence infer, with extreme probability, that such long streamers as Abbe saw last July, Myer in August 1869, Feilitsch in June 1860, and several Swedish observers during the eclipse of 1733, are produced in the same way as comets' tails, and therefore really extend (as they seem to do) radially from the sun. It is also certain that if they did not extend radially from the sun, their always seeming to do so would be altogether inexplicable. So that the theory to which we are led in one direction leads us also out of what would else be a very perplexing difficulty in another direction.

Some recent inquiries which I have made have led me to the belief that the radial streamers are after all meteor streams, opposite streamers (in the cases cited above) being parts of the same meteor streams lying (at the time) beyond the sun, so as to be seen athwart his disc.

METEORIC ASTRONOMY.

THE views respecting meteors now held by astronomers are of such extreme importance, whether viewed directly, or regarded in relation to the inferences which seem to flow from them, that they may be regarded as affecting our ideas respecting the present constitution as well as the past history and the future fate of all the orbs which people space. I propose to consider the position to which meteoric astronomy has at present been brought, and to point out the connection between the results now established, and the subject of the solar corona, which has recently occupied so large a share of the attention of astronomers.

We need not consider here the history of the earlier meteoric theories. It would be difficult to show that the mere correct ideas of the Greek and some of the Roman writers who have spoken of meteors were less purely speculative than the later view that meteors are mere phenomena of our own air, like lightning or the aurora. And although the theory that meteors are bodies which have been expelled from lunar or planetary volcanoes was discussed by mathematicians of eminence, yet it was not based on exact observation; so that the calculations of Laplace, Obbers, and others, serve rather to illustrate the skill of those mathematicians than to establish any conclusions of value.

We may begin, then, by considering the first important fact tending to prove the extra-terrestrial nature of meteors and shooting-stars-the circumstance, namely, that meteoric displays occur commonly on certain days of the year.

It had been noticed in very early times that a display of shooting-stars nearly always occurs on the night of August IO. This being known in calendars as St. Lawrence's Day, the meteors which fall on that day have been called the tears of St. Lawrence. Among astronomers, however, they are more commonly called Perseides, for a reason presently to be cited. Not so early, but still many years before the true theory of meteors began to be recognised, it was known that on or about the 12th or 13th of November shooting stars are commonly seen. When Humboldt, after witnessing the remarkable display of 1799, invited special attention to this circumstance, ancient records were examined, and it was found that for several centuries this particular part of the year had been characterised by star-showers. 'Time out of mind,' says Sir John Herschel, 'those identical nights more often, but sometimes those immediately adjacent, have been habitually signalised by such exhibitions.'

It cannot be too often insisted upon-since doubts are frequently expressed respecting the truth of the modern. theory of meteors-that this circumstance of periodicity suffices of itself to demonstrate the extra-terrestrial nature of these objects. There are no meteorological phenomena which recur persistently on August 10 and on November 13; terrestrial volcanoes are not then exceptionally active; the moon on those dates may be in any part whatever of her orbit. The one circumstance to which phenomenal recurring on a particular date can be held to point is the recurrent passage by the earth of a particular part of her orbit on that date. If we picture the earth circling around the sun in her wide orbit, once in each year, and remember that year after year as she crosses the particular point corresponding to August 10, and again as she crosses the particular point corresponding to November 13, her air is alive (as it were) with meteors, we at once see that this is because she comes across the meteors at those stages of her circuit. It is precisely as though a person who travelled