Clayton — ^7 -Day Auroral Period and the Moon. 87 



near new moon, and in the mean of each consecutive five the 

 maximum number of auroras occurs exactly at the time of last 

 quarter. The irregularities in the totals are almost entirely 

 smoothed out in the column giving the mean of five, and even 

 after they are thus smoothed the number of auroras near the 

 last quarter of the moon is about ten times as great as the 

 number reported near full moon. The results in columns 3 

 and 4 are shown in the diagram, fig. 2, the lighter broken line 

 representing the unsmoothed totals and the heavy line the 

 mean of five. This great variation in the number of auroras 

 according to the age of the moon furnishes a clue to the 

 reversal of the time of maximum frequency of auroras in the 

 27-day period, as a result of which the maximum occurs near 

 the time of the moon's greatest northern declination in sum- 

 mer and near the time of the moon's greatest southern decli- 

 nation in winter. In the summer the greatest southern decli- 

 nation of the moon occurs near full moon, so that the chief 

 maximum of auroras occurring at that time is obscured. In 

 the winter the greatest northern declination occurs near 

 full moon, so that the secondary maximum at that time 

 is obscured, and the chief maximum at the time of great- 

 est southern declination is well marked. But as yet there 

 appears no reason why there should be a secondary maximum 

 of auroras at the time of the moon's greatest northern declina- 

 tion, for according to Ekholm's theory of the moon's induc- 

 tion of electricity on the earth, the difference in potential 

 between the earth and the air ought to be at a minimum at 

 the time of the moon's greatest northern declination. 



It is well known that when the sun crosses the plane of the 

 earth's equator there is a marked maximum in the number of 

 auroras. This is well shown for the United States by the fol- 

 lowing table, prepared by Professor Hazen, showing the num- 

 ber of stations reporting auroras during each of the twelve 

 months of the year for an interval of 23 years : 



Jan. Feb. Mar. April May June July Aug. Sept. Oct. Nov. Dec. 

 1005 1455 1396 172Jf 1270 1061 1223 1210 1735 1630 1240 912 



The explanation of this, according to the theory of Mr. E. B. 

 Elliot (Bull. Phil. Soc, vol. i, p. 45), is that the number of 

 auroras depends on the rate at which the earth, in its annual 

 orbit, cuts through cosmic lines of electric force. 



Since Dr. Ekholm's theory fails to explain the secondary 

 maximum shown by the curves at the time of the moon's 

 greatest northern declination and the tertiary maxima at the 

 time the moon crosses the plane of the earth's equator, I am 

 led to believe that some magnetic effect must be involved, 

 probably engendered by the unequal cutting of the lines of 

 magnetic force by the moon in its motion around the earth. 



