203 



< 



«arth. By reason of the sun's rotation the sun-spot advances, 

 culminates, and then recedes ; such cases would be expected 

 to be connected with terrestrial aurorae exhibiting the same 

 phases. Should a sun-spot be suddenly created on the surface 

 of the sun directly facing the earth, auroral phenomena would 

 appear without advance indications. Examples such as are 

 required by the above regular movements of sun-spots are 

 commonly witnessed. 



Following up the argument of sun-spot bombardment, 

 we should expect a recurrence of auroral phenomena at an 

 interval pf the rotation period of the sun. 



On looking over the journal, it appears that in some cases 

 this does happen to be the case. For instance, the aurora 

 of April 27 was followed by one — a sun's rotation period later 

 in May, and again, though somewhat less prominently, in 

 June. Following the bright aurorae of the early days of 

 June was a demonstration- — a sun's rotation period later at 

 the end of the month. That of May 23 may be regarded as 

 appearing again on June 18. There appears, therefore, to 

 be some ground for supposing these to be repetitions; at any 

 rate, the suggestion should not be hastily dismissed, for at 

 the best such a cycle would not be likely to be prominently 

 evidenced in scanning an auroral record. There are many 

 circumstances which would exclude the uossibilitv of recording' 

 the reappearance of auroral phenomena connected with a 

 particular sun-spot after an interval of a sun's rotation 

 period. 



Chief amongst these is the fact that only a proportion of 

 sun-spots continue to exist for the time necessary, and even 

 those may be too feeble to produce auroral phenomena. 



Secondly, co-operating circumstances may not be favour- 

 able. Again, the meteorological conditions may be unfavour- 

 able for observing. In the latter case, as the relation 

 between auroral phenomena and earth magnetism is not yet 

 finalized, it would be unsafe to rely upon magnetic records. 



4. Diurnal Variation. 



Distribution of Auroral Phenomena in Relation to Time 

 of Day. — Account is taken only of the months June, July, 

 and August, because of their more complete record. Text 

 iig. 2 shows graphically the relative average amount of visible 

 aurora seen at Cape Royds throughout the twenty-four hptors 

 of the day (see text fig. 1). 



The flattening at noon is at least partly due to the 

 obscuration of faint effects bv twilight. 



The maximum daily aurora took place between 2.45 a.m. 

 and 3.0 a.m. During the afternoon, between 4 p.m. and 



