varying from 2/1 to 3/1, the former ratio being characteristic of the hours oi less 

 frequent occurrence, and the latter ratio characteristic of the hours of more frequent 

 occurrence of aurora. This difference simply expresses, therefore, the fact that aurora 

 is not only most likely to be seen at certain hours, but is also of greater extent at 

 the same hours. This agrees with what has been stated before, viz., that the 

 manifestations of aurora appear to originate in a definite quarter, being seen low 

 down on the horizon at hours of minimum frequency and minimum display, approaching 

 the station (increasing in altitude) as the hour of maximum display and maximum 

 frequency approaches. 



In addition to the principal maximum at 4 hours, indications appear of a secondary 

 maximum at 16 or 17 hours, followed by a minimum at 19 hours. The minimum at 

 12-13 hours is, of course, due to twilight, local noon being at about 13 hours. That 

 the minimum in the late afternoon is a real one can hardly be doubted, when one 

 considers that, at the equally dark hour in the morning corresponding to (say) 18 hours, 

 aurorae are about six times as likely to be seen as at the latter hour.* 



Though the number of observations is small at any particular hour, it is to be 

 remarked that the observations of Shackleton's Expedition are in fairly close agreement 

 with ours. Thus Mawson records a principal maximum at 3 hours, a minimum at 

 17-20 hours, and a secondary maximum at 16-17 hours. Mawson's observations differ 

 from those of the present Expedition, however, in that auroree were less frequently seen 

 in the late morning than in the corresponding early afternoon hours (i.e. hours equally 

 distant from local noon). However, the observations of the two expeditions are not 

 strictly comparable, as apart from the different methods of observation, Mawson's 

 results refer only to the months of June, July and August. 



In addition, the local surroundings (hills and mountains) of any observing station 

 must affect the distribution to some extent, and at neither station was a clear view 

 afforded to the horizon. 



The observations by Bernacchi on the previous Scott Expedition, whose winter 

 quarters were also not far removed from the present Expedition, indicated in 1903 

 a maximum at 2 hours (local time) with a subsidiary maximum at 16 hours and a 

 succeeding minimum at 18 hours. In view of the restricted field of view from that 

 station the agreement is very good. It is of great interest, however, to note that in 

 1902 no trace was observable of the secondary maximum in the afternoon. 



7. DIURNAL FREQUENCY OF DIFFERENT TYPES OF AURORA. 



Aurora of 70 altitude or over. 



Analysis of the whole of the observations in 1911 (85 cases) when aurora was seen 

 from Cape Evans at an altitude of 70 or over, shows a well-marked diurnal frequency 



* See footnote to preceding page. If the organized watch had included observations at 18 hours, 

 this disproportionality would have been reduced to some extent. In support of the reality of the 

 secondary maximum, it can lie stated that the maximum is not the result of -more frequent observation 

 of the sky during this period. 



14 



