1895.] on Atmospheric Electricity, 475 



from observation if the aurora was always accompanied by discbarges 

 through the earth's surface. Except in the polar regions, these 

 auroras do not seem to affect the normal fall of potential. There is a 

 third view we may take as to the circulation of electric currents indi- 

 cated by the aurora : the return current may take place in space out- 

 side the earth's atmosphere. A good deal might be said in favour of 

 this view, and the rotation of the earth's magnetic field in space 

 might be a sufficient cause for the production of these currents ; but 

 this is not the place to enter further into this question. 



Calculations made from observation on the height of the aurora 

 have generally resulted in an altitude of from 100 to 200 miles, 

 except in the polar regions, where the aurora seems occasionally to 

 descend to a much lower level. It has also been noticed that auroras 

 are associated with certain bands of cirrus clouds, and this seems to 

 indicate that although the luminous phenomenon is sufficiently 

 intense to be observed at only great heights, yet the electric pheno- 

 mena may descend to the level of the cirrus. 



As regards the connection between the aurora and the sunspot 

 period, further observations in the polar regions are needed. On the 

 one hand, we have Paulsen's * statement, derived from observations in 

 Greenland, to the effect that the greatest number of auroras are seen 

 when sunspots are at their minimum, that is, at a time when in our 

 own latitudes the number is smallest ; and, on the other hand, we 

 have Nordenskiold's observations, which seem to point in the opposite 

 direction. In a publication which contains much important matter 

 on the geographical distribution and form of the aurora borealis, 

 Nordenskiold contrasts the appearances he has observed in the Vega 

 during the winter of 1878-79, passed in the Behring Straits, with 

 that previously observed in 1872-73 to the north of Spitzbergen. 

 According to this author, the auroras, during the minimum sunspot 

 period in 1878-79, were " hardly worthy of his notice by the side of 

 those observed in 1872-73." But although only faintly luminous, 

 the auroras of 1879 were persistent and regular in shape. They did 

 not affect the magnetic field, and seem to show a regular and con- 

 tinuous, though weak, electric discharge. The arc and streamers in 

 1872 were much more brilliant and much more irregular. Some 

 objection may be raised against these observations, in so far as they 

 refer to different places, and local circumstances may have affected 

 the phenomenon ; but in the face of the very careful description he 

 gives us, we cannot as yet accept Paulsen's results without further 

 confirmation. 



The problem of atmospheric electricity, like that of terrestrial 

 magnetism, presents special features in the arctic regions, and until 

 we possess a greater number of observations in those little accessible 



* Paulsen, ' Danske Videnskab. Selskabs Forband.,' 1889. (I have not seen 

 tbe original memoir, but only an abstract in the ' Jahrbuch der Astronomie und 

 Geophysik,' 1890.) 



