The figures in the last column are found by dividing the average of auroras in 

 each month, by the number of hours less one, from sunset to sunrise at Toronto 

 (taken as a middle latitude), on the 15th of that month ; they show in a striking 

 manner the diminished frequency of the phsenomenon about the winter solstice, 

 and its great development at the vernal equinox. The returns for the last three 

 months of 1851 are not all collected. 



Tlie stations of observation in the &st and second groups 

 are not yet numerous enough to decide the question, whether 

 the aurora ever appears in the exterior when it is absent in the 

 interior circles ; but in forming an opinion on the number which 

 extend from the interior to the exterior of them, we must not 

 forget, that, notwithstanding the large number of obsei-vers, 

 both regular and occasional, in the third group, and their M'ide 

 distribution, a considerable proportion of the entries in Table III. 

 rest at present upon an observation at only one station*; and 

 unless particulars are given, which is unfortunately not always 

 the case, may be reasonably regarded as doubtful. Where par- 

 ticulars are given, there can be of course no doubt. 



The observations made under the direction of the Smithsonian 

 Institution begin to be general in March 1849; and the stations 

 are so numerous, that we ought perhaps to consider observations 

 to have been possible every night. Table III. has been made as 

 complete as possible, by including some observations kindly 



* Of the total number of 261 obsei-vations in 1850, 64 are at one station 

 only J of the total number of 207 in 1851, 71 arc at present at one station 

 only — the majority of these in the third group. The proportion to which 

 any doubt ciin attach is not large. 



