320 MEMOIRS OF THE AMERICAN ACADEMY. 



By the help of my second catalogue, and the supplement, the yearly march in the 

 number of auroras observed, from month to month, can be accurately determined for 

 Sitka, Toronto, Quebec, Newfoundland, etc., for which the materials in the first cata- 

 logue were insufficient. But the large addition of more than two thousand to the 

 aggregate number of auroras will throw little light on our inquiry into the secular 

 periodicity of the aurora, inasmuch as they belong, for the most part, to a short 

 period of years, and furnish no long series of observations at any single place. And 

 even those subordinate fluctuations, which some suppose to move synchronously 

 with the disturbances on the sun's surface, require for their elucidation continuous 

 observations, at the same place, for many decades. 



If sufficient evidence exists for believing that the disturbances in the sun's atmos- 

 phere, which reveal the solar spots, are associated in time with those other disturb- 

 ances, whatever they may be, which result in displays of the aurora, this coincidence 

 between the two classes of phenomena, pointing to a mutual dependence or to a 

 common origin, will guide speculation, if it do not even furnish the key for the 

 unlocking of some of the great mysteries of the physical world. The testimony in 

 favor of such a coincidence must, therefore, be subjected to a rigid examination. A 

 glance at the curves which represent the periodicity of these phenomena will satisfy 

 any one that the march of the changes in spot-frequency is far more regular, both in 

 degree and in time, than that which is furnished by the aurora. 



I shall now examine separately each of the twenty-eight maxima and minima of 

 the solar spot curve between 1700 and 1860. 1. First comes a maximum in 1705. 

 A decided maximum of the aurora is found in 1707. If the secular period of the 

 aurora is taken at fifty-seven years, a secular minimum would have occurred about 

 the year 1701. This may account for the postponement of the subordinate maximum 

 by two years. Six auroras were observed in Great Britain in 1709, a few before, but 

 none afterwards for six years. 2. The spot-minimum of 1712 is accompanied by 

 an aurora-minimum. I attach less weight, in general, to the agreement of minima, 

 because the evidence is negative. The number of observed auroras is subject to 

 violent transitions, from year to year, not witnessed in the number of solar spots. As 

 few auroras were observed in 1714 and 1715 as in 1712. Nevertheless the year 1712 

 is the central period of half a dozen years during which auroras were infrequent. No 

 auroras were recorded in Germany in 1712, 1713, 1714, and 1715, although four or 

 five in all were seen during the four preceding years, and more afterwards. 3. Next 

 comes the spot-maximum of 1717. There seems to have been an aurora-maximum in 

 Paris. But an error exists undoubtedly in the record, which has doubled the number 



