position of the spiral in space (due to variation in the magnetic forces), causing the 

 aurora to be seen overhead at a time when it is normally far north of the station. 



/ 



Quite possibly the two are associated through a common cause. 



It is necessary to examine more closely the relation between mean magnetic 

 character number and type of aurora, and, with this object, the mean magnetic 

 character numbers for hours when aurorae were seen close to the zenith and in the 

 sectors where they w r ere least frequently observed have been calculated from Dr. Chree's 

 figures, together with the corresponding number for all observations of " clear, no 

 aurora," at the exact hours. The figures for 1911 are as follows : 



Auroras in the zenith .. .. Mean character number, 0-88. 



Auroree in the sectors where they 



were infrequently observed .. ,, ,, ,, 0-87. 



" Clear, no aurora " .. ,, ,, ,, 0-75. 



These three numbers differ little from one another and from the general means 

 and confirm the conclusion that the position of an aurora has little, if any, effect on 

 the mean magnetic character number. The figure corresponding to the large number 

 of occurrences of " clear, no aurora," confirms our conclusion that the correlation 

 between magnetic character number and aurorse is not close except for brilliant, 

 coloured or quickly moving forms. 



The fact that the position of the aurora bears little relation to the magnetic 

 character number is clearly proved by the figures given by Dr. Chree for the Cape 

 Adare auroral observations. These figures are reproduced below for four classes of 

 aurorso arranged in descending order of brightness, the corresponding magnetic character 

 numbers being taken from the Cape Evans magnetic traces some hundreds of miles 

 distant. 



Cape Adare, 1911. Class I. Mean magnetic character number (Cape Evans) 1-32. 



j; ** a ); ' '22. 



III. : , 1-00. 



IV. 0-87. 



Mean character number, May- August, 0-84. 



It will be seen that there appears to be a closer correspondence between auroral 

 intensity at Cape Adare and magnetic character number at Cape Kvans than was the 

 case with auroral observations at Cape Evans. This increased correspondence is, 

 however, probably accounted for by the more favourable meteorological conditions 

 and by the fact that the aurora at Cape Adare was of such intensity thai the presence 

 of moonlight was insufficient to render observations impossible. 



35 c 2 



