306 METEOROLOGY AND ALLIED SUBJECTS. 



north of the line of greatest frequeDcy toward the Pole the frequency 

 and magnitude of the aurora diminishes more rapidly than toward the 

 Equator, although for no place in that zone is the aurora entirely in- 

 visible, as is the case in lower latitudes. 



(2) With regard to the direction in which auroras are seen, after dis- 

 cussing a large number of observations, the author finds that the lines 

 of greatest frequency and neutral directions are, in all probability, sub- 

 ject to periodical changes in frequency and extent, whereby they soon 

 come nearer the Pole and again depart from it, as, in fact, could have 

 already been suspected from the different altitudes and positions of the 

 observed segments, arcs, &c. 



(3) As regards the distribution of auroras in space, the author con- 

 cludes that the great part of the auroras has no great extension, or the 

 phenomenon is produced by conditions which are very local in their 

 nature. 



(4) In regard to the altitude of the aurora above the earth's surface, 

 the author states he can make no advance on the conclusions previously 

 formulated by others. The altitudes above the earth's surface at whiclt 

 polar lights develop are very various, and the altitudes of these regions, 

 at least for the lower limit, diminish with the latitudes. 



(5) The extent of the aurora and the duration vary within wide limits. 

 The greatest extent in latitude has been from 29° south up to 82<=' 

 north, and the greatest in longitude has been about 280^. The greatest 

 duration has, apparently, been from August 28 to September 7 [1859]. 



(6) With regard to diurnal period, he finds that the aurora attains 

 only one maximum and one minimum; the former is usually about 10 

 p. m., the jirincipal exception being at Point Barrow, where the maxi- 

 mum would appear to be at 3 or 4 p. m. 



(7) In regard to the annual period, nearly all the series of observa- 

 tions show two maxima and minima. 



(8) In regard to the eleven year and the secular periods, it would 

 appear that the periodicity of the aurora can be determined more accu- 

 rately in proportion as we know the perodicity of the sun spots. 



(9) The following table shows the comparison between ajiparent sun 

 spots and auroras, so far as both of these data can be gathered from im- 

 perfect records : 



Dates of maxima and minima of ohserred number of auroras. 



