whole period of observation) on which the brighter and fainter aurorse were seen at 

 different times of day. In brackets are also shown the corresponding figures for the 

 period May 26th to July 31st. The third column gives the number of occasions on 

 which " Clear, no aurora," was logged at exact even hours, during the shorter period. 



From the Table it will be seen that, for the period May 26th- July 31st, the 

 brighter aurora? show the early morning and evening maxima quite well (2-4 hours 

 and 18-20 hours, respectively), the minimum at 20-22 hours being not well marked. 

 The situation is quite otherwise with auroras of lesser brightness, which are fairly 

 evenly distributed over the dark period of the day, a distribution somewhat similar 

 to that of " Glow " aurora?. 



It will further be observed that the inclusion of days in March, April, May, August, 

 and September, does not modify the distribution to any great extent. 



It might seem from consideration of these figures that the brighter aurora? occur 

 most frequently at times when the aurora is close to the zenith (4 hours), a result 

 which is reversed at the Cape Evans station (compare Table 8). Though Table 15 

 shows almost equal numbers of aurorae in the morning and evening periods, this 

 is far from true for the aurorse classed as " bright " and " very bright," three 

 times as many being recorded in the evening as in the morning period. This shows 

 that the brightest aurora? rarely do occur at Cape Adare at times when aurorae are 

 most frequently seen near the zenith, a result in direct opposition to that deduced 

 from a division of the aurora into only two classes of brightness. ' On the other hand, 

 very faint aurorse are most likely to be seen in the morning hours. 



Another interesting result of closer analysis of the available figures shows that 

 bright aurora? are very rare in the late morning, while faint aurora? are comparatively 

 very common. This result one would be inclined to explain as due to an incorrect 

 estimate of the influence of the twilight arch in the late morning. If this were the 

 case, however, it would be expected that a similar effect would be shown in the 

 corresponding early hours of the afternoon. For the class of very faint aurorse, this 

 is not the case, which lends strength to the contention that there is a real difference 

 between the diurnal variation of bright and faint aurorse. 



As in the case of the observations at Cape Evans, the ratio C/(A-f-B) (in Table 15) 

 varies during the day from 0-1 between 2 and 4 hours, to 1-1 at 16 hours. 



28 



