

1863.] 531 



1852, 1853, 1854, 1855, 1856, 1857. On sheet III. all the obser- 

 vations at each nominal hour through all the months January from 

 1841 to 1847 are combined to form the January curve; all those 

 through the months February to form the February curve, and so 

 on. On sheet IV. similar month-curves are formed from the period 

 1848 to 1857. It is remarked that the origin of coordinates neces- 

 sarily represents the mean declination and mean horizontal force in 

 each month. 



The author then points out that the means for each month are 

 themselves subject to an annual inequality, which can be ascertained 

 with little difficulty. The values of these inequalities are exhibited, 

 for declination and horizontal force, separately for the period 1841- 

 1847 and for the period 1848-1857; those in the first period far 

 exceed in magnitude those in the second (as holds also with regard 

 to all the diurnal inequalities). 



If we wished to exhibit the hourly state of magnetism, as referred 

 to the mean state given by the supposition of uniform secular change 

 of normal magnetism, we ought to apply these quantities with sign 

 changed, to the origin of coordinates in each curve, in order to form 

 a new origin of coordinates. For the year-curves, the numbers de- 

 stroy each other, and no new origin of coordinates is produced ; for 

 the month-curves, however, they shift the origin materially. The 

 author does not perceive that any facility for theoretical reference or 

 other advantage is gained by this step. 



On examining the year-curves, it is seen that from 1841 to 1848 

 their magnitude very slowly increases, with a small change of form, 

 but from 1848 to 1857 their magnitude very rapidly diminishes, with 

 a great change of form. Some great cosmical change seems to have 

 come upon the earth, particularly affecting terrestrial magnetism. 

 On comparing these year-curves with the month-curves, especially 

 with those for the period 1848-1857, it appears that the change of 

 the year-curves from 1848 to 1857 nearly resembles that of the 

 month-curves from summer to winter ; and the author points out as 

 a possible step to a physical explanation of the change from 1848 to 

 1857, that the magnetic action of the sun upon the earth's southern 

 hemisphere may have remained nearly unaltered, while that on the 

 northern hemisphere may have undergone a great diminution. 



The author then alludes to the curves representing the hourly 



