82 Clayton — %7-Day Auroral Period and the Moon. 



dental arrangement is calculable but is complicated by the fact 

 that there were auroral displays on other days than those in 

 the series. It is sufficient for my purpose to show that the 

 probability of chance occurrence in series of this kind is so 

 small that the more reasonable conclusion is that there is some 

 natural cause for this grouping, and this, I think, is in evidence. 

 These sequences of auroras at intervals of about 27 days 

 have been noticed before, and have heretofore been attributed 

 to a connection of the aurora with the rotation period of the 

 sun. Fritz in Europe, and Dr. Veeder and Professor Bigelow 

 in this country have each maintained the existence of such a 

 connection, though they differ somewhat in regard to the 

 length of the period. The length of the lunar period formed 

 by the oscillations of the moon to the north and to the south 

 of the earth's equator is also about 27 days, and I was led to 

 investigate the connection of this period with that of the 

 auroras, especially as there seems a physical reason for such a 

 connection. Dr. JNils Ekholm and Svante Arrhenius have 

 shown (Ueber den Einfluss des Mondes auf den electrischen 

 Zustand der Erde, Stockholm, 1894) that the difference of elec- 

 trical potential between earth and air at the stations in northern 

 Europe is greatest when the moon is south of the equator and 

 least when the moon is north of the equator, while the reverse 

 is true at Cape Horn. Their explanation is that the moon is 

 an electrified body, charged negatively like the earth, and that 

 when it is north of the equator it acts more directly by induc- 

 tion on the parts of the earth beneath, diminishing the electri- 

 cal potential in the northern hemisphere between the earth and 

 the air, while the potential increases in the southern hemi- 

 sphere. The reverse takes place when the moon is south of 

 the equator. For the purpose of comparing the positions of 

 the moon north and south of the equator with the occurrence 

 of auroras, I selected from the Nautical Almanac (for the data 

 from which I am indebted to Professor E. C. Pickering) all 

 the dates when the moon reached its greatest northern declina- 

 tion. Taking these as the first day of each period, I classified 

 the numbers of auroras observed in the United States into 

 periods of 27 days. The auroral data was taken from the 

 American Meteorological Journal, vol. ii, p. 227, from the 

 Monthly Weather Keview, and from the Proceedings of the 

 Rochester Academy of Science. Table II shows the number of 

 stations in the United States reporting auroras on each day of 

 the lunar period. Columns 1 to 6 show the totals in groups of 

 fifty periods ; column 7 shows the totals on each day of the 

 lunar period for the entire interval of 23 years from 1870 to 

 1892 ; column 8 gives the departures of the totals of each day 

 from the mean of the column. The results show a very 



