98 
Proceedings of the Royal Society 
in the side of a copper pot, one of which was filled with oil and 
heated by means of a brass Bunsen, while the other was filled with 
water at the temperature of the air of the room. The temperatures 
of both ends of the magnet could thus be ascertained by means of 
mercurial thermometers. 
In the same line as the magnet and on both sides of it were two 
small magnets. These were cemented to the backs of small con- 
cave mirrors, suspended by single silk fibres, and placed in glass 
cases to guard them against currents of air. The deflections of 
the small magnets were measured exactly as in the reflecting 
galvanometer; and from the nature of the arrangement, and the 
important fact made out from these experiments, viz., that even 
when the poles of a magnet are at different temperatures the mag- 
netic strength is the same in each, it follows that the absolute 
magnetic strength in either pole of the large magnet is directly as 
the tangent of the angle of deflection of the contiguous small 
one, and, therefore, will be measured by the reading on the cor- 
responding scale. 
y* d + y)* * 
Te 
/ m m \ 
^v7"(T+yjv 
( m m \ 
x* ~ (T + xf)^ 
/ m m, > 
\~x 2 ~(L + x) 2 J 
NS are the poles of the fixed magnet, m its absolute magnetism. 
N a = x, N a 1 = y, NS = Z. The couples indicated are those pro- 
duced by the large magnet and the earth’s magnetism, E, on the 
small magnets. 
For any deflection <9, if the lengths of the small magnets be 
negligible compared with x and y, we have 
E sin 0 = m ( 
” Q + 
| cos 0 . 
\ m a tan 6 . 
E sin 6 = m j 
r 1 
1 ^ 
1 cos 6 . 
m a tan 0. 
(J+y) V 
