astronomer. Kepler had observed it fully 70° in length a day or two previous. The famous comet of 1680, which was attentively watched by Newton and Halley, was attended by a train 90° long. That observed in the southern hemisphere in 1689, had a tail more than 60° in length, and was two hours and a half in rising; and in another which appeared six years afterwards, the tail was nearly of equal magnitude. The comet of 1769 was attended by a long and conspicuous tail, but observers differ considerably in their estimations of its apparent length, even on the same day, which can only be attributed to the state of our atmosphere at the various stations. This circumstance had been previously noticed in reference to the comet of 1680, but it was more marked in the present case. Thus on the 9th of September, 1769, Maskelyne at Greenwich considered the length of the tail 43°; at Paris it was judged to be 55°; at the Isle of Bourbon, it was traced 60° and more by La Nux, while Pingré saw it fully 75° long, being at sea at the time, between Teneriffe and Cadiz. Two days afterwards, this astronomer found it 90° long, while La Nux traced it 97° from the nucleus. The only comet of the present century which has been accompanied by a tail of very great length, is that of 1843, its average extent was about 45°, but on one occasion a narrow ray of light attained the enormous distance of 65° from the head.

It would lead us far beyond the limits of the present work, were we to particularize all the varied phenomena which have been observed in the tails of comets, but there is one singular appearance in the

trains of great comets which we must not pass over in silence. It consists of apparent vibrations or coruscations, similar to the pulsations peculiar to the Aurora Borealis. These vibrations commence at the head, and appear to traverse the whole length of the tail in a few seconds of time. The cause was long supposed to be connected with the nature of the comet itself, but Olbers pointed out that such appearances could only be attributed to the effects of our own atmosphere. The reason is this: the various portions of the tail of a large comet must often be situated at widely different distances from the earth, so that it will frequently happen that light would require several minutes longer to reach us from the extremity of the tail than from the end near the nucleus. Hence, if the coruscation were caused by some electrical emanation from the head of the comet, travelling along the tail, even if it occupied only one second in passing over the whole distance, several minutes must necessarily elapse before we could see it reach the end of the tail. This is contrary to observation, the pulsations being almost instantaneous.

Gregory of Tours mentions a comet in January 582, the train of which resembled the smoke of a distant conflagration, a description which may perhaps have some relation to these vibrations or coruscations. The first distinct mention of the phenomenon is to be found, we believe, in the Chinese Annals in reference to the comet of 615, the tail of which was between 50° and 60° long, and during the night had a kind of libration to and fro. Kepler says the tail of the comet of 1607 was short one moment, but extended itself in the

twinkling of an eye. Longomontanus, in describing the immense tail of the comet of 1618, states that it had an enormous vibration,' and Father Cysat says it appeared as though it had been agitated by the wind; the same phenomenon was remarked by Kepler, Wendelin, and other observers of this splendid comet. Hevelius noticed similar undulations in the tails of the comets of 1652 and 1661, and Pingré says he had observed them in the train of that which was visible in 1769. The pulsations were very distinctly seen in the tail of the grand comet of March 1843, and they have been remarked in a greater or less degree in other

cases.

Another curious phenomenon occasionally observed in the tails of comets is a curvature of the extremity, or of a greater or less portion, so as to give the whole train the form of a sabre. The ancient historians frequently make use of this simile in describing the aspect of these bodies. The comet which appeared about the time of the Battle of Salamina, A.D. 479, was of this form, as also those observed at Constantinople in 912, 1340, 1402, 1456, and many others. The Chinese compare the tail of the comet of 1232 to the tusk of an elephant. The grand comets of 1264, 1618, and 1689, exhibited curved tails 80° or 100° in length, the latter being particularly described by an observer in the southern hemisphere as having a striking resemblance to a great sabre.' Pingré noticed a curvature in the tail of the comet of 1769, the convexity towards the north, and at times a second. small arc was formed near the extremity, turned in the opposite direction; the same appearances were

remarked by La Nux at the Isle of Bourbon. The train of the last great comet in 1843 was very sensibly bent downwards towards the horizon at the end of March, when it became conspicuous in this country, and the smaller comet of December and January 1844-5 exhibited the same phenomenon.

CHAPTER II.

OF THE REAL DIMENSIONS OF COMETS-PHASES OBSERVED IN SOME OF THEM-THEIR PHYSICAL CONSTITUTION CHANCES OF COLLISION WITH THE EARTH.

HAVING described the apparent dimensions of various comets we must now say a few words relative to their real magnitude. When the distance of one of these bodies is known, and we have observed the angular diameter subtended by the nucleus or head, it becomes a very easy matter to ascertain the true diameter in the same way that we find the dimensions of the sun, moon, and planets. But as it never happens that the borders of the nebulosity are sharply defined, and but rarely so as regards the nucleus, our determinations of their real dimensions are necessarily open to a good deal of uncertainty. One thing, however, is quite certain, that the cometic atmosphere surrounding the nucleus varies greatly in extent in different comets, and even in the same body at different epochs during its visibility. The actual length of the tail of a comet may be computed by trigonometry, when we know the distance of the nucleus from the earth and its position in respect to the sun, always assuming that