ATMOSPHERE IN RELATION TO HUMAN LIFE AND HEALTH. 311 



The aurora appears to be tlie result of the agitation and vibration of 

 particles of air under the influence of the passage of an electric current, 

 diverging from the magnetic polar regions. The current passes where 

 the resisting power is least, that is, in highly rarefied air, dense air 

 and a vacuum both offering too much resistance to be used for the 

 course of the current. It strongly affects telegraph wires and corre- 

 sponds with earth currents of uncommon intensity. It has been sup- 

 posed by Sabine and others to be connected with disturbances in the 

 sun, which, again, depend on the position of the planets. Sun spots 

 and auroras were considered to be at a maximum in periods of eleven 

 years; aurora? and earth currents to be due to small but rapid changes 

 in the earth's magnetism; the upper conducting strata of the air to 

 behave like a secondary coil, and the sun to act like a primary current 

 which produces magnetic changes in the core of a Euhmkorlf machine. 

 There seems to be no doubt of a connection between the periods of sun 

 spots, of the variation of the magnetic needle, and of aurora?. 



Some observers have noted a connection between these lights and 

 great cyclonic storms, but they are certainly not always followed by 

 bad weather, and in North America have been associated with clear 

 skies. Moreover, the height at which they traverse the air renders it 

 unlikely that they should be either the cause or effect of disturbances 

 in the lower air. 



Occasionally the elevation of moisture and cirrus cloud to a great 

 height may afford a readier than ordinary means of transit to electric 

 currents. Generally, however, cirrus cloud does not extend to one- 

 tenth of the calculated height of the aurora, and can hardly aid in 

 forming a passage for the current. That some visible effect of induc- 

 tion may be produced on cirrus and high cirro-cumulous, which are 

 themselves electrified, is not improbable. The present writer was once 

 greatly struck by a very extraordinary arrangement of high cirrus and 

 cirro-cumulus clouds in closely packed, detached, reticulated, and nearly 

 rectangular compartments, covering the whole area of the sky overhead, 

 from 9 to 9.50 a. m. on November 17, 1882, in London, and learned after- 

 wards that at about 10 a. m. a great magnetic storm had occurred over 

 the country. The radiant point was about north. The appearance 

 of the clouds was represented on paper at the time, and the diagrams 

 were afterwards submitted to members of the Eoyal Meteorological 

 Society. 



The simultaneous appearance of an aurora in northern Europe and 

 in America rather discounts the supposed connection between this phe- 

 nomenon and the weather, for changes very rarely take place about the 

 same time and in connection with each other over this wide area. 

 March and October, the months of maximum display, happen to be 

 months which are often windy in England. The cause of the aurora 

 is rather to be sought in changes which come within the scope of astro- 

 nomical inquiry. The spectroscope has not given much information 

 regarding the nature of the substances which emit the light. The 



