TKRRESTKIAL MAGNETISM. 



131 



result of the slow secular change would take several years to 

 produce, may occur and be reversed many times during a few 

 hours of a magnetic storm. The disturbed condition may 

 extend over one or more days. The intensity and irregularity of 

 the disturbance is greatest in the high-latitude zones where 

 auroral phenomena are most frequent. 



There is a close connection between magnetic storms and solar 

 disturbance, storms being frequent when the sun's surface is 

 much spotted, and vice versa. They are consequently most 

 numerous at times of sunspot maximum. Again, the sun 

 rotates on its axis in a period which, as \dewed from the earth, 

 is between 27 and 28 days. Mr. Maunder has shown that there 

 is a marked tendency for magnetic storms to recur after one or 

 more such intervals of time. This indicates the sun as the source 

 of the initiating factor in the production of magnetic storms, 

 and suggests that this factor is communicated to the earth along 

 narrow streams proceeding from, and apparently rotating with, 

 the sun. As I conceive it, the action is somewhat as follows : 

 The sun ejects electric particles from disturbed regions on its 

 surface, and these particles travel outwards in limited streams 

 which, on account of the great speed of ejection, seem to revolve 

 with the sun. When such a stream happens to be directed 

 towards the earth, many of the particles are precipitated into 

 the earth's atmosphere, their paths being deflected to some extent, 

 however, by the earth's own magnetic force. The bombardment 

 of the upper atmosphere by these particles produces another 

 ionized layer, like that due to the ultra-violet light from the sun, 

 except that the former extends all over the earth. This layer, 

 however, is not only ionized but electrically charged, and the 

 entangled charges, when thus brought to comparative rest in 

 our atmosphere, exercise a strong mutual repulsion. This results 

 in an upward expansion of the air (the action resembling that 

 of an electrically charged soap bubble, which, owing to the same 

 cause, expands until either it bursts or the tension balances the 

 repulsion). In this way the charge is enabled to escape, perhaps 

 carrying away with it a portion of the atmosphere itself. Some 

 of the most characteristic features of magnetic storms are 

 explicable as consequences of these actions. 



Before concluding, I would add a few words relating to 

 magnetic observation. This is usually di^dded under two heads, 

 observations of the first kind comprising those of which the 

 object is to determine the value of one of the elements of magnetic 



K 2 



