AURORA BOREALIS. 215 



tion? of the lieavens, and sometimes almost tlic entire celestial vault. Sometimes 

 its light but little exceeds that of the milky way. Its light is generally faint, 

 especially in the upjjer part of the sky, but sometimes its accumulation near the 

 horizon produces a pretty intense light resembling a vast conflagration. The 

 great disparity between the light of auroral vapor when viewed near the zenith 

 and near the horizon is a ])roof that its ver^cal thickness is small in comparisoa 

 with its horizontal dimensions. Sometimes this distant light resembles the 

 twilight. This auroral vapor may appear during any phase of a grand aurora, 

 and is probably due to an extrem-^ diffusion of the substance of auroral arches. 

 It is frequently seen during the intervals between the disappearance and reap- 

 pearance of arches and beams. 



Colors of the aurora. — The color of the aurora is ordinarily white; some- 

 times it is of a pale yellow, and occasionally it becomes reddish. When sthe 

 movements of the auroral beams become rapid, the yeHuw color flows from the 

 extremities of the beam towards the centre, while one of its extremities bncomes 

 red and the other green. The red of the aurora is usually tinged with violet, 

 and does not correspond to the red of the prismatic spectrum. The green is 

 tolerably pure, blended, perhaps, with a slightly bluish tint. The red tint is 

 the brighter of the two, and that which disappears the last. Sometimes the 

 red and green are arranged parallel to the length of the beam, but more fre- 

 quently the lower part of the beam is red and the upper part green. Sometimes 

 the entire auroral illumination becomes of a red color. The, coloring of the 

 aurora is mo-t frequently observed between 10 and II o'clock, and this is 

 usually the period of greatest brilliancy of the aurora. 



Geographical dhirihution of auroras. — Auroras are very unequally distributed 

 over the earth's surface. They occur most frequently in the higher latitudes, 

 and are almost unknown within the tropics. At Havana (latitude 23^ 9') but 

 six auroras have been recorded within a hundred years; and south of Havana, 

 auroras are still more unfrequent. As we travel northward from Cuba, auroras 

 increase in frequency and brilliancy; they rise higher in the heavens, and oftener 

 attain the zenith. In order to determine the law of distribution of auroras over 

 the northern hemisphere I have collected observations from 128 localities, 

 showing, as far as possible, the average number of auroras seen annually at 

 each place.* These observations are not as complete as could be desired, tiv- 

 quently comprehending a period of less than a year; nevertheless, when we 

 project them upon a chart we find them unexpectedly consistent with each other. 

 If we travel from the equator northward along the meridian of Washington, 

 we find on an average, near the parallel of 40^, only 10 auroras annually. 

 Near the parallel of 42'^, the average number is 20 annually; near \-j\ 

 the number is 40; and near the parallel of 50^, it amounts to SO annually. 

 Between this point and the parallel of 62^ auroras are seen almost every 

 night. They appear high in the heavens, and as often to the south as the north. 

 Further north tliey are seldom seen except in the south, and froin this point 

 they, diminish in frequency and brilliancy as we advance towards the pole. 

 Beyond latitude 62", the average number of auroras is reduced to 40 annually. 

 Beyond latitude 67", it is further reduced to 20; and near latitude 78^, to 10, 

 annually. If we make a like comparison for the meridian of St. Petersburgh, 

 we shall find a similar result, except that the auroral region is situated iurther 

 northward than it is in America; the region of SO auroras annually, being found 

 between the parallels of 66" and 75". 



Upon the accompanying chart the dark shade indicates the region where the 

 average number of auroras annually amounts to at least 80; and the lighter 

 shade indicates the region where the average number of auroras annually 



'American Journal of Science, n. s., v. 30, pp. 89-90. 



