162 EH Loomis on Solar Spots, Magnetic Declination, ete. 
whether auroral displays exhibit a periodicity in any given lo- 
ceality or region, and for this purpose we desire a long series of 
faithful observations at a single station or a limited number of 
stations ; or if satisfactory observations of this kind cannot be 
obtained, then we must employ observations from a limited re- 
gion of country where the records have been the most complete 
and continuous. 
servations at St. Petersburg (Smith. Rep. 1865, p. 227) are con- 
tinuous from 1726 to 1811; and the observations at Berlin (Heis 
Wochenschrift, Jan. 1870, p. 47) are continuous from 1700 to 
1800. I have combined these observations with those at New 
Haven and Boston, and the correspondence of the resulting num- 
bers with the fluctuations of the solar spots is quite satisfactory. 
Inasmuch, however, as I have not been able to obtain a con- 
tinuation of the Petersburg and Berlin observations to the present 
time, I have thought it best to extend the area of observation, and 
include all published observations from Europe south of the par- 
allel of 55°. By this method of comparison we eliminate to a 
considerable extent the anomalies introduced into the obser- 
ead catalogue of European auroras classified by parallels of 
atitude, and I have used this as my principal basis in the sub- 
auroras recorded in Europe south of the parallel of 55°; column 
third the auroras at New Haven and Boston ; colunin fourth the 
bers thus resulting are given in column : 
For a few years preceding and following 1787, column _— 
auroras sufficient to excite a suspicion that these numbers are 
relatively too great. The Palatine Meteorological Society, which 
