110 Major-General Sabine on the Cosmical Features 



sponding to each month, showing the mean or normal posi- 

 tions at each hour. When these monthly Tables are compared, 

 the results are found to arrange themselves into two distinct 

 categories, the one comprising the six months in which the sun 

 is in the northern signs of the zodiac, and the other the six 

 months in which he is in the southern signs. The difference of 

 the diurnal variation shown by the two categories is of a very 

 marked character, and is always the same in whatever part of 

 the globe the station of observation is situated. It is obviously 

 a periodical affection common to the whole globe, having a year as 

 its cycle, with semiannual epochs coinciding as nearly as may 

 be with the equinoxes. A semiannual variation of similar cha- 

 racter is shown in each of the three magnetic elements; but, as 

 the facts of the declination have been more extensively examined 

 than those of the two other elements, I have confined myself to 

 them in Plate II. which illustrates this particular case of the 

 solar action*. In this Plate the several hours of local astrono- 

 mical time at each of the stations are shown by the figures at the 

 top and bottom. The unbroken curves in each figure correspond 

 to the months from April to September, when the sun is in the 

 northern signs; and the broken curves to the months from 

 October to March, when he is in the southern signs. The 

 divergence of the unbroken and broken curves respectively from 

 the mean annual solar-diurnal variation at the station (which for 

 this purpose is represented in each of the figures by a horizontal 

 line) shows the characters of the semiannual irregularity in 

 the two opposite categories. The Plate includes stations in each 



* Plate II. Semiannual inequality : — 



Fig. 1. At Kew, from hourly positions of the declinometer magnet from 

 June 1, 1858, to December 31, 1861, inclusive; omitting disturbed 

 positions equalling or exceeding a difference of 3'*3 from the normal 

 of the month and hour. 



Fig. 2. At Toronto, from hourly observations from 1843 to 1848 inclu- 

 sive ; omitting disturbed observations equalling or exceeding a dif- 

 ference of 3' - 6. 



Fig. 3. At Nertchinsk, from hourly observations from 1851 to 185/ inclu- 

 sive ; omitting disturbed observations equalling or exceeding a dif- 

 ference of V96. 



Fig. 4. At Pekin, from hourly observations from January 1, 1852, to 

 October 31, 1855, inclusive; omitting disturbed observations equal- 

 ling or exceeding a difference of 2' '25. 



Fig. 5. At St. Helena, from hourly observations from September 1, 1842, 

 to August 31, 1847, inclusive; omitting disturbed observations 

 equalling or exceeding a difference of l'*78. 



Fig. 6. At the Cape of Good Hope, from hourly observations from 

 October 1, 1841, to September 30, 1846, inclusive; omitting: dis- 

 turbed observations equalling or exceeding a difference of 1''88. 



Fig. 7- At Hobarton, from hourly observations from January 1, 1841, 

 to September 30, 1848, inclusive; omitting disturbed observations 

 equalling or exceeding a difference of 2'' 13. 



