136 Dr. C. Chree on the Law 
momentum has been kept as constant as possible.” If this 
explanation be correct, we should infer, even from the annual 
-Means, a variation in the p of the mirror magnet amounting 
to nearly :04. 
According to the records of the Royal Alfred Observatory, 
Mauritius, the mean annual values of P! for two English 
magnets 24, and 24, changed from —5:059 in 1890 to 
— 4°236 in 18938, and then to —5:006 in 1897. This might 
be accounted for by a change of little over ‘01 in the p of one 
of the magnets. ; 
Fis: 1. 
O 
S N 
x 
§ 24. In general, the ideal conditions of the deflexion 
experiment cannot be strictly fulfilled. The magnets are not 
only of finite section, but their axes may be neither strictly 
horizontal nor perpendicular to one another, while their 
centres may differ in level. Having already, § 15, considered 
the finiteness of the cross section, we shall in treating the 
other defects suppose the magnets linear. Also, as a small 
inclination in the deflected magnet reduces practically in the 
same proportion the couples acting on it in the horizontal 
plane, whether due to the deflecting magnet or the earth’s. 
field, we shall, to avoid undue complexity, suppose this 
magnet strictly horizontal.. In fig. 1—which is practically 
Borgen’s* figure—N’‘S’ represents the axis of this magnet, 
C its centre. The deflecting magnet if properly adjusted has 
its axis horizontal when suspended, but unless this axis 
happens to coincide with the mechanical! axis it will ordinarily 
in the deflexion experiment be shghtly inclined to the 
horizontal. In fig. 1, NS (middle point O) is the projection 
of the magnetic axis of the. deflecting magnet on the hori- 
* Archiv der Deutschen Seewarte, 1891, no. 2, p. 2. 
