122 AURORA BOREALIS pletely across the sky." The position of the luminous region is not known. Arago was of opinion that each observer sees his own aurora, somewhat as each observer of a rainbow sees the luminous arc differently placed. Sir John Herschel says " no one can doubt that the light of the aurora originates nowhere but in the place where it is seen." But it has been con- sidered that the most favorable conditions for the determination of the height of auroral gleams are presented when the auroral corona is formed. Now this corona always surrounds the point toward which the magnetic dip- ping needle points. Yet the magnetic dipping needles at different stations are not directed toward one and the same point ; so that what- ever the auroral corona may be, it does not seem to hold a definite place, in such sort that its distance can be determined by simultaneous observations ; for it is the essential principle of the method of simultaneous observations that the lines of sight should be directed to one and the same point. Nor is it easy, on Herschel's theory, to interpret the fact that the auroral corona has been seen at stations distant more than 1,000 miles from each other, and always around the part of the heavens pointed to by the magnetic dipping needle. For a point im- mediately overhead at one station, and 100 miles from the earth's surface, would be be- low the horizon of a station 1,000 miles dis- tant. We seem forced to adopt the conclusion that though there is no analogy whatever be- tween the aurora and the rainbow, yet Arago was right when he asserted his belief that as each observer sees his own rainbow, so each observer sees a different aurora. We should thus be led to consider whether the nature of the luminous emanations the direction, for in- stance, of the luminous flashes composing them may not explain the formation of the auroral corona. In this case the position of the observer would affect the appearance of the phenomenon. If we assume that reliance can be placed on the observations by means of which the height of the auroral arch has been estimated, we must assign a considerable elevation to many of these lights. On Oct. 17, 1819, an aurora was observed simultaneously at Gosport, Kes- wick, and Newtown Stewart, in Great Britain ; and from the calculations made by Dalton the meteorologist, the arch was estimated to be 101 or 102 miles above the earth. More re- cently Sir John Herschel estimated that the arch in the aurora of March 9, 1861, was 83 miles above the earth. But he remarks that "the auroral light has been seen below the clouds, as in the polar seas by Parry, Sherer, and Ross, on Jan. 27, 1825 ; near the chain of the Rocky mountains on Dec. 2, 1850, by liar- disty ; and at Alford in Scotland on Feb. 24, 1842, by Farquharson ; nay, even habitually seen as if hovering over the Coreen hills in the last-mentioned neighborhood, at a height of from 4,000 to 6,000 miles." Herr Galle, from observations made during the aurora of Feb. 4, 1872, estimates the height of the auroral corona on that occasion at 2fi5 miles above the sea level. Prof. Olmsted's conclusion that the auroral arch is seldom below 70 miles in height or above 160 miles, would thus appear to be negatived. But probably all such estimates must be abandoned, and "our meteorological catalogues," as Arago advised, " must be disen- cumbered of a multitude of determinations of height, though due to such great names as Mai- ran, Halley, Krafll, Cavendish, and Dalton. "- The extent of the earth's surface over which the same aurora has been visible has some- times been remarkable. Kamtz mentions that on Jan. 5, 1769, a splendid aurora was seen simultaneously in France and in Pennsylvania ; and that the remarkable aurora of Jan. 7, 1831, was seen from all parts of central and northern Europe, in Canada, and in the northern parts of the United States. But even these instances, and others of the same kind which might be cited, are surpassed in interest by the circum- stance that auroras of great brilliancy occur simultaneously over the major part of both the northern and southern hemispheres. Kamtz mentions that when Capt. Cook's observations are analyzed, it appears that on every occasion when he observed an aurora 'australis an aurora borealis had been seen in Europe, or else the agitation of the magnetic needle proved that around the northern magnetic pole an auroral display must have been in progress. The aurora of Feb. 4, 1872, was seen not only in America and Europe, and over the northern hemisphere generally, as far S. as lat. 14 N., but in Mauritius, in South Africa, in Australia, and probably over the greater part of the south- ern hemisphere (for Mauritius is much further north than southern auroras are ordinarily seen). Mairan and Cassini were the first to point out that auroras do not occur at all times with equal frequency or in equal splendor. The former mentions that a great number of auroras were seen at the beginning of the IGth century (a misprint probably for the 17th, as the con- text seems to imply) to beyond the year 1624, after which nothing more was heard of them till 1686. Kamtz mentions that between 1707 and 1790 there was a remarkable increase fol- lowed by decrease of auroral action, the max- imum frequency being attained in 1790. Prof. Olmsted considered that there was sufficient evi- dence to establish a period of 20 years during which auroral displays are frequent, precede " and followed by intervals of from 60 to 65 year during which few are witnessed. But it is open to question whether the existence of this long period is as yet established. The actual fri quency of auroras cannot be inferred from ob- servations made in temperate latitudes, where alone hitherto any attempt has been made to determine long periods. The longest riod which lias been thoroughly established one of about 11 years. This period is associ- ated with the occurrence of magnetic disturb- ances in cycles of 11 years. The connection
