Notwithstanding this variability in intensity and form from minute to minute. 

 it is quite evident that the incidence of aurorse, both from hour to hour and from clay 

 to day, is governed by more or less definite laws. Thus, if the aurora is of unusual 

 brightness during one hour of the day it is probable that the brightness will be 

 above normal during the greater part of the same day. In other words, if a bright 

 aurora is seen at one hour, the aurora will probably still be bright in succeeding hours 

 of the same day. 



Not only is this the case, but it may be taken as a general rule that a " disturbed " 

 day will more likely be followed by another disturbed day than by a " quiet " day. 

 This tendency for disturbed conditions to continue for more than one day may be 

 readily seen from Table 1, which gives the days in 1911 which have been estimated 



TABLE 1. List of Disturbed Days, Cape Evans, 1911. 



April 28. 



May 1, 2, 3. 



May 6, 7, 8, 9. 



May 15, 16. 



May 19. 



May 21, 22. 



June 5, 6. 7, 8. 



June 22, 23, 21, 25. 



June 28, 29, 30 (mornings only). 



July 1, 2, 3. 

 July 19, 20, 21. 

 July 28, 29, 30, 31. 

 August 5. 

 August 7. 

 August 26. 

 August 29. 

 September 3. 

 September 7. 

 September 13, 14. 

 September 16, 17, 18, 19. 

 September 22. 



as " disturbed," i.e. days of bright auroral display. From the Table, it will be seen 

 that " runs " of several days are not uncommon, and there can be little doubt that 

 these runs would have been even more pronounced if the meteorological conditions 

 had been more favourable. 



The converse of these cases appears to be equally true that is, a quiet hour is 

 likely to be followed by a quiet hour, and a quiet day by another quiet day. 



Too much stress should, however, not be laid on the tendency indicated in 

 Table 1 for the aurora to continue for several days at a time, as the days noted 

 comprise a goodly proportion of all days when the moon was below the horizon. 



By allotting to each day numbers from to 4 according to the estimated auroral 

 intensity shown on that day, extracting the days allotted numbers greater than 2, 

 and comparing with the mean number for the preceding and following days, an 

 estimate can be made of the tendency towards recurrence of strong aurorse on 

 successive days. 



The mean numbers for days n 5 to n+5, where n is the clay showing intensity 

 greater than 2, is given below and illustrated in fig. 1 : 



u 5 n 4 n 3 n 2 it. I . n+1 n [-2 +3 n-f-4 n+5 

 1-47 1-47 2-12 2-14 2-62 3-27 2-35 2-00 1-98 1-34 1-62 



A4 



