No corroboratory evidence can be offered on this point as a result of the present 

 observations. Though in about half the number of cases there was a more or less 

 regular rotation (with the sun) of the arches, the trend of the arches remained fixed 

 in many other cases. Only two observations were, however, recorded in which there 

 was slight rotation in the reverse direction, both these cases having been immediately 

 preceded by a slight rotation in the usual direction. Some evidence on this point 

 can be drawn, however, from Table 3. From this Table, it will be seen that there is 

 a tendency for arches with any given trend to be grouped about a more or less definite 

 time, though variations from this mean time up to 6 hours are not unknown. The 

 average time corresponding to various trends of arches are given below : 



S.W.-N.E. 



S.S.W.-N.N.E. 

 S.-N. 



S.S.E.-N.N.W. 

 S.E.-N.W. 



1 hr. 20. 

 1 hr. 20. 

 3 hr. 30. 

 5 hr. 00. 

 o hr. 00. 



There is, therefore, some evidence for a rotation of the arch in the sense of the 

 apparent movement of the sun, but, on the other hand, there seems to be no indication 

 whatever of a tendency for a given trend to recur twice a day, at intervals of 12 hours. 

 (Owing to the twilight arch, it is hardly to be expected that a double periodicity would 

 be indicated, except in the case of arches running from N.W. to S.E. at 5 hours, at 

 which time observations are possible at the same hour in the morning and the 



afternoon.) 



5. DISTRIBUTION OF AURORA IN AZIMUTH. 



As indicated in Table 4, auroras show a strong preference for certain sectors and 

 aversion for others. In this Table, column 2 gives the number of occasions on which 

 aurorse were recorded in the sector North to North-East, column 3 the number of 

 occasions recorded in the sector North-East to East, and so on. The period referred 

 to is from April 24th to September 30th, 1911, inclusive, observations at exact hours 

 alone being counted. 



TABLE 4. Distribution of Aurora in Azimuth, Cape Evans, 1911. 



11 



