436 



question. Now, if a plane is laid through the sun's axis parallel 

 with the axis of the earth, this plane intersects the ecliptic in two 

 points, whose longitudes are 286 and 106, and has an inclination 

 of 83'2, both axes forming an angle of 25' 8. Accordingly, the 

 northerly prolongation of both axes will converge (as in 1 and 2) 



Earth. Sun. *" . Earth, 



when the longitude is 286, or about seventeen days after the summer 

 solstice, and the southerly prolongation (as in 2 and 3) when the 

 longitude of the earth is 106, or sixteen days after the winter 

 solstice. About sixteen days after the equinoxes the axes are in the 

 position 2, and the radius vector forms the greatest angle with the 

 above-named plane, viz. 83'2. It is therefore evident, that the 

 greatest magnetic induction takes place at the two solstices, but of 

 opposite character, as the north poles converge in one, and the south 

 poles in the other semiannual period : the change takes place six- 

 teen days after the equinoxes, exactly as you have found by observa- 

 tion, and I regard this circumstance as very important evidence for 

 my hypothesis, although you have shown that it could also be ac- 

 counted for by the fact, that all magnetic induction takes some time 

 ere it attains its maximum.' 



" If the sun's magnetic axis does not coincide with the axis of ro- 

 tation, then I suppose we shall find, by more minute examination of 

 the observations, that there exists also a small magnetic variation 

 corresponding to the sun's geocentric rotation, or 27'68 days. 



" There seems also to be strong reason to suppose that Hansteen's 

 discovery about the annual periodic frequency of the aurora borealis, 

 which has a marked maximum at the equinoxes, and even a more 

 marked maximum at the solstices, is connected with the same cause." 



