34 ON THE RECENT SECULAR PERIOD 



throws much doubt and uncertainty upon the accuracy of the observation. Finally, 

 if the case be as supposed by Captain Parry, it will only be an exception, and a 

 very rare one to a general rule, and will by no means prove that ordinary auroras 

 are at a lower level than that assigned them by Dalton, Twining, and others, who 

 have estimated their heights from undoubted parallaxes, on the sure principles of 

 trigonometry. 1 



XI. Periodicity. — Auroras of the higher classes have three distinct forms ofjjcriodi- 

 city — a diurnal periodicity, commencing, arriving at their maximum, and ending 

 at definite hours of the night; an annual periodicity, rarely or never occurring in 

 June, and the greatest number of the higher order clustering together in September 

 and November, these last bearing a striking resemblance to each other; and a secular 

 periodicity, the most remarkable of all, recurring in great series. 



1. Diurnal periodicity. — We have already seen that auroras of the higher orders 

 usually begin near the end of evening twilight, come to their maximum from 10 to 

 11 o'clock, and more frequently a little before 11, and last nearly or quite through 

 the night, although those of an inferior kind usually terminate before midnight. 



2. Annual periodicity. — For the period of twenty years, beginning with 1832 

 and ending with 1851, of 51 auroras of the first and second classes, 14 were in 

 September, 8 in November, 8 in August, and 6 in March ; while there were none 

 in June and October, but 1 in December, and but 2 in January and February, 

 respectively. The sum total for August, September, and November, was 30, 

 averaging 10 to each month, while for December, and January, and February, the 

 number was only 5, averaging but If to each. If, however, we regard number 

 merely without reference to intensity, then for the period of seventeen years, from 

 1832 to 1818 inclusive, out of 780 auroras the sum for each month was as follows : 



1. April 93 7. May 63 



2. September 84 8. November G2 



3. August 82 9. January 58 



4. October GO 10. February 57 



5. July GG 11. June 46 

 G. March 65 12. December 38 



In respect both to number and intensity, June and December stand at the bottom 

 of the scale, while September stands highest in intensity and next to the highest in 

 number. October is the fourth in respect to number, but at a minimum for 

 intensity. These results, however, do not appear to be uniform for different 

 periods. According to Celsius, out of 384 auroras, observed in the North of 



1 An instance communicated to Captain Lefroy by Mr. Ilardisty, showing the low descent of the 

 aurora, is stated as follows : It (the aurora) appeared between me and the trees on the opposite side of 

 the river, which could not have been 40 feet above the level of the stream, the trees toward the top of the 

 hill being high above it. — Lefroy's Second Report, 1850-1, p. 14. A similar case is mentioned by Rev. 

 Mr. Cowles in the American Journal of Science, XIII, p. 429. But should it be proved that, in 

 some rare instances, portions of auroral vapor have descended near the earth, this fact will not militate 

 against our views of its cosraical origin. 



