OF THE AURORA BOREALIS. 11 



range of 1° 40', while its ordinary diurnal variation is not more than four minutes. 

 The thermometer, at 11 o'clock, was at 33°; it shortly fell to 31°, and continued 

 nearly at this point during the remainder of the night, a degree of cold consider- 

 ably below that of the few preceding nights. About 3 o'clock, the sky grew cloudy, 

 "and the moon rising shortly afterwards, further observations were prevented; but 

 the continued disturbance of the magnetic needle, would induce the belief that the 

 aurora prevailed through the night, and to a late hour next morning." 



On the evening following the above exhibition of the aurora borealis, that is, on 

 the 18th of November, 1835, a fine aurora was witnessed at London, having all 

 the characteristics of the first class except the corona. It is described in the 

 Arcana of Art and Science, for 1836, by Dr. Armstrong, as seen at Vauxhall. The 

 writer remarks that the aurora had made its appearance, with greater or less dis- 

 tinctness, on every clear and calm evening from the middle of October; a frequency 

 very unusual for that climate. 



On the 11th of December, of the same year, an aurora of the more brilliant kind 

 (2d class) was observed by Captain Bonnycastle, at Toronto, which he has described 

 and figured in the thirtieth volume of the American Journal of Science. 



The Regents Eejwrt, for 1836, contains an elaborate paper by Professor Joslin, 

 then of Union College, Schenectady, on the auroras of 1835, and several preceding 

 years. He had carefully observed and recorded, during the five years ending in 

 1835, fifty-six auroras, of which twelve were of the last year; and he was the first, 

 in this country, to describe certain appearances, seen in the daytime, resembling 

 auroral exhibitions, and which, according to him, are truly such. Dr. Joslin's paper 

 also contains valuable records of the state of the weather preceding and following 

 an aurora, from which he deduces the general rule that, "previous to an aurora, the 

 barometer is rising and the thermometer falling; and that the air is much nearer 

 than usual to the point of saturation with moisture." 1 



It may be remarked that each of these circumstances betokens that fair and 

 transparent state of the atmosphere, which is peculiarly favorable to the exhibition 

 of the aurora. 



Dr. Gibbons, of Wilmington, in the State of Delaware (lat. 39° 41'), has given 

 a record of all the auroras seen by him during the six years from 1827 to 1833, 

 amounting to 52, — a number which is evidently much greater than the average for 

 that latitude, and, therefore, plainly indicating that the time fell within the secular 



1 In Jameson's Journal we find the following account of an aurora borealis seen in the daytime, at 

 Canonmills, September 9th, 1827. The morning of Sunday, September 9th, was rainy, with a light 

 gale from the N.E. Before mid-day the wind began to veer to the west, and the clouds in the north- 

 western horizon cleared away; the blue sky in that quarter assumed the form of the segment of a very 

 large circle, with a well-defined line ; the clouds above continuing dense, and covering the rest of the 

 heavens. The centre of the azure arch gradually inclined more to the north, and reached an elevation 

 of nearly 20°. In a short time very thin fleecy clouds began to rise from the horizon within the blue 

 arch; and through these, very faint, perpendicular streaks of a sort of milky light could be perceived 

 shooting ; the eye, being thus guided, could likewise detect the same pale streaks passing over the 

 intense azure arch ; but they were extremely slight and evanescent. Between nine and ten in the even- 

 ing of the same day, the aurora borealis was very brilliant-. 



