are rarely seen at this station. Reference to the Cape Evans observations shows that 

 in both cases the most likely trend for arches is in a direction at right angles to the 

 azimuth in which aurora? are least frequently seen. 



Again no definite evidence can be offered as to whether the arches show a 

 rotation with the sun or no. A scrutiny of the results shows simply that the rotation, 

 if existent, is more likely to be with the sun than in the opposite sense. There appears 

 to be little tendency for the arches to exhibit the same trend twice daily, as postulated 

 by Mawson for the Ross Island aurorse, except possibly in the case of the arches showing 

 a N.W. to S.E. trend. 



5. DISTRIBUTION OF AURORA IN AZIMUTH. 



Table 10 has been drawn up to illustrate the distribution of aurora both in 

 azimuth and time, the totals for all hours of the day being shown in the last line. 

 The period covered by this Table is May 26th to July 31st, observations only at the 

 regular times of observation being included. From the Table and from fig. 4, it will 

 be observed that there is a well-marked variation in azimuth, the azimuth of 



TABLE 10. Distribution of Aurora in Azimuth and Time, Cape Adare, 1911. 



minimum frequency of aurora being slightly to the south of south-west. These figures 

 are very closely proportional to the corresponding figures for a period May 26th to 

 June 30th, the relative order in frequency of the various sectors being the same in 

 both cases. As in the case of the Cape Evans observations, the " maximum " and 

 " minimum " azimuths are consistently located in nearly the same sectors at all 

 hours, except for the minimum azimuth at 6 hours, which is in the east. This hour is, 



24 



