224 



AURORA BOREALIS. 



January . . . 

 February . . 



March 



April 



May 



June 



.Jnly 



August.. .. 

 September^ 

 October . . . 

 November . 

 December .. 



Canada. 



Sum. 



16 

 31 

 24 

 38 

 22 

 17 

 21 

 14 

 19 

 27 

 26 

 16 



173 

 210 

 240 

 267 

 191 

 179 

 244 

 238 

 293 

 236 

 215 



l.:9 



If we classify tlie entire series of observations by seasons, we shall have in — 



Spring . 

 Summer 

 Autumn 

 Winter. 



Boston and 

 New Haven. 



300 

 308 

 357 

 257 



New York. I Canada. 



314 

 301 

 315 



222 



84 

 52 

 72 

 63 



Sum. 



698 

 661 

 744 

 542 



These observations show a decided minimum in December, and there is appa- 

 rently another minimum in June. There are apparently two maxima — one in 

 April and the other in September. The slight diminution in the number of 

 auroras in summer as compared with the spring may be ascribed, at least in part, 

 to the longer continuance of daylight. If we increase the number of auroras 

 recorded for the month of June, in the ratio of the number of hours in the two 

 months upon which ordinary auroras might be seen, we shall conclude that 

 auroras are nearly as abundant in June as in April. We infer, then, that there 

 is a very decided dirahuition in the frequency of auroras in December, and a 

 period of maximum frequency from April to September, wiili perhaps a slight 

 diminution during the intervening month of June. 



Secular periodicity of auroras. — The number of auroras seen on different 

 years is extremely variable. Sometimes, for several years, auroras are remarka- 

 ble for their number and magnificence, and then succeeds a barren interval 

 during which auroras are almost entirely forgotten. 



If we compare the observations made at any one station for a long period of 

 years, we shall discover not merely an inequality in the number of auroras upon 

 Buccessive years, but this inequality bears a strong resemblance to a secular 

 periodicity. This is shown iu a long series of observations made at Boston and 

 New H.iven. 'J'he observations for Boston have been published by Professor 

 Lovering in the Memoirs of the American Academy, vol. ix, pp. 101-120. The 

 observations for New Uaven have been recently pub'ished by the Connecticut 

 Academy of Arts aud Sciences.* The combined series of obs.erv;ition3 extends 

 from 1742 to 18o4, and embraces in the aggregate 1,222 auroras, not counting 

 duplicates. The following table exhibits the number of auroras observed each 

 year at New Haven and Boston, according to this combined series of observa- 

 tions : 



« t 



'Transactions of the Conn. Acad of Arts and Sciences, vol. 1, pp. 9-172. 



