Chemistry and Physics. 311 



in June (about 5th) and December (about 6th), neither pole is 

 turned towards or away from the earth, but occupies an inter- 

 mediate position. Hence we see that the equinoxes occur in the 

 same months as those in w T hich one or other of the solar poles is 

 turned towards the earth, while the neutral positions of the 

 solar poles in relation to the earth occur in the same months as 

 the solstices. 



The accompanying diagram [here omitted] shows graphically the 

 relation between the two curves representing the variation of the 

 sun's declination and the change of the latitude of the sun's 

 center or the variation of the amount of the tilt of the solar 

 poles, in relation to the earth throughout a year. 



It will be seen that the curve representing the tilt of the solar 

 axis is nearly (a little less than) a quarter of a phase in advance 

 of that indicating the declination change, so that the maximum 

 or minimum point of the latter curve is only slightly in advance 

 of the mean points respectively of the former curve. 



If, therefore, these solar polar regions are capable of disturb- 

 ing the magnetic and electric conditions on the* earth, as has 

 been above suggested, then, when they are most directed to her 

 at the equinoxes, the greatest effects during a year should be 

 recorded, and when they are least directed the effects should be 

 at a minimum. 



It will not be necessary here to refer at any great length to 

 statistics relating to the annual inequality of magnetic disturb- 

 ances and auroras, for these have been very efficiently worked 

 out and the results published by Mr. William Ellis.* 



Mr. Ellis has shown that the curves of frequency of magnetic 

 disturbances at Greenwich and Paris are very similar, " showing 

 maxima at or near the equinoxes, and minima at or near the 

 solstices." These also, he further points out, are similar, with 

 regard to the epochs of maxima, to the curve representing the 

 frequency of the aurora at London. In the case of auroras 

 observed in Edinburgh, Northeast Scotland and in different 

 regions in Scandinavia, the months in which the greatest fre- 

 quency is recorded are September and October (perhaps more 

 generally October) and March and April (perhaps more gener- 

 ally March). Mr. Ellis is inclined to the opinion that there 

 is a small tendency for the autumn maximum to become a little 

 later (from September to October) and the spring maximum 

 somewhat earlier (from April to March) as higher latitudes are 

 approached. 



Further, he points out that in more northern latitudes the 

 mid-winter minimum of lower latitudes appears to diminish and 

 eventually disappears, so that the curve of frequency of the 

 aurora between October and March is practically flat with a 

 small intermediate maximum about January. This change in 

 form of the frequency curve in regions in close proximity to the 



* Monthly Notices, E.A.S., vol. lx, p. 142; vol. lxi, p. 537; vol. lxiv, p. 

 229. 



