very unlike those through which the earth herself passes. The meteor systems required by the theory must be much denser and much more brightly illuminated than the August and November systems. To say they must be much more brightly illuminated is equivalent to saying that they must be much nearer the sun. And in this we see an escape from another difficulty. Meteor systems very near the sun would be far more likely to appear as streamers extending radially from him than systems at a great distance from him. A distant system might, by a mere chance, so appear. For instance, if a total eclipse of the sun had occurred on or about May 10, 1865, the November meteor system (whose richest part was then crossing the earth's track at the point she occupies on November 13th) would have appeared, if discernible at all, as a streak athwart the sun's place in the sky, and therefore forming two rays on opposite sides of him, somewhat 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 13th. At no other time in the year except November 13th and May 10th, or about these dates, could the November system present such an appearance. But a system traveling 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 strewed 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 color," he said, "was white, or resembling that of Venus; there was no flickering or unsteadiness; it was not separated
