194 MAGNETIC DISTUEBANCES AND AURORA. 



The absence of auroral records from the remainder of the year means nothing more than " too much 

 daylight," to use Mr. BERNACCHI'S words. 



Even in the Midwinter months there are gaps in the auroral record : in 1902, from April 16 to May 5, 

 from May 15 to 30, from June 16 to 29, from July 14 to August 4 ; in 1903, from April 12 to 18, from 

 May 8 to 15, from June 7 to 13, from July 7 to 11, and from July 31 to August 10, all inclusive. Of the 

 third of these gaps Mr. BERNACCHI says: "From June 15 to June 30 bright moonlight or overcast skies 

 prevented any aurora being seen," and, presumably, the explanation of the other gaps is similar, as the 

 dates of full Moon were, in 1902, April 23, May 23, June 21, July 20, and, in 1903, April 13, May 12, 

 June 11, July 10, and August 8. 



111. It will be observed that many more auroras were seen in 1903 than in 1902. I am informed by 

 Mi\ BERNACCHI that this is not due, at least in any large measure, to difference in the observational 

 methods or increased activity in the observers. It is thus, presumably, a true physical phenomenon. 

 1902 was near sun-spot minimum, with a WOLFER'S frequency of only 5-0, while the frequency for 1903 

 was 24 4. In northern temperate latitudes auroral frequency normally increases with sun-spot frequency 

 [it is, however, doubtful whether the same is true to the north of the zone of maximum auroral frequency], 

 thus a difference between 1902 and 1903 is not surprising. The difference between the two years seems 

 due partly to the greater length in 1902 of the intervals in which no auroras were observed. This rather 

 suggests that the cause was difference in intensity rather than anything else. In 1903 the average aurora 

 may have been more intense than in 1902, and so have suffered less from the causes tending to render it 

 invisible. Some collateral evidence of this is afforded by a consideration of the number of separate entries 

 of aurora in Mr. BERNACCHI'S list. 



On some days aurora is noted at one hour only, but on most days, when it is recorded at all, at several 

 hours. 



There is a certain amount of overlapping, so that there may be a trifling error in my estimate of the 

 number of separate entries. The figures show, however, a very large difference between the two years, 

 there being 250 separate entries for 1903 as against 125 for 1902. The ratio 20 : 10 between these two 

 numbers is substantially larger than the ratio 17 : 10 between the number of days of aurora in the two 

 years. This obviously supports the view that in 1903 the average aurora retained for a longer time the 

 intensity necessary to render it visible, from which a higher maximum intensity would naturally be 

 inferred. 



If this is a main cause of the excess of visible auroras in 1903, then the preceding figures are, at least, 

 strongly suggestive of the view that the phenomena whose visible side is aurora were seldom, if ever, 

 wholly absent on Midwinter days at Winter Quarters. Thus the fact that the magnets were practically 

 never undisturbed for five minutes at a time cannot safely be interpreted as evidence of the continued 

 presence of some cause of irregular magnetic disturbance other than that associated with the seat of 

 aurora. 



Another consideration should be borne in mind. Assuming, what few people now doubt to be true, 

 that aurora is the visible manifestation of electrical action in the atmosphere, the existence of a very 

 bright auroral band or streamer may mean an electrical current of unusually high intensity reckoned per 

 unit of cross-section, but one having a comparatively small section. Thus the existence of visible aurora 

 may mean only local concentration and no great total quantity of current. Thus the ultimate causes of 

 aurora and magnetic disturbance may be the same, without any close parallelism being exhibited between 

 the apparent intensities of the two phenomena. 



112. Two attempts were made to trace the possible interconnection of aurora and magnetic disturbance. 

 The character of the magnetic curves was considered at all the times when aurora was noted, and a rough 

 judgment passed. The magnetic trace was characterised as " quiet," " normal," " moderately disturbed," 

 and " somewhat highly disturbed." These terms may be thus interpreted : " Normal " means that I 

 regarded the amount of disturbance as about average ; " quiet " means that distinctly less than the 

 average amount of disturbance existed ; " moderately disturbed " means somewhat above the average 

 amount of disturbance ; and " somewhat highly disturbed " that the existence of more than usual 

 disturbance obtruded itself even on casual inspection. The average amount of magnetic disturbance 



