545 
of Edinhurgli, Session 1883 - 84 . 
of tlie frame from an outside stand, so that tlie frame can be moved 
round or taken out of its place without disturbing the magneto- 
meter. The displacement of the mirror is observed, as usual, by 
the motion of the reflected image of an illuminated slit or wire. 
The whole is shown in fig. 1. 
To determine the magnetic meridian — 
1. Place the frame in a vertical plane, with the magnetometer at 
its centre, and its length approximately at right angles to the 
magnetic meridian. 
2. Observe the position of the reflected imago on tlm magneto- 
meter scale before closing the circuit. 
3. Make the circuit so that the field due to it has the same sign 
as the terrestrial field. The image will in general be displaced, 
and its movements will be quickened on account of the increase of 
directive force. Turn the frame until the image comes back to its 
original position. 
4. Eeverse the current: unless the adjustment in operation 3 
has been correct, the image will again be displaced. The current 
is to be regulated (by the resistance coil) to prevent the equilibrium 
from being unstable or too nearly neutral. Turn the frame, if need 
be, until the image comes to its original position. hTow draw a 
line on the table, using one edge of the frame as a straight-edge, or 
in some other way note its position. 
5. Break the current ; remove the frame and replace it inverted, 
and with its former east end now pointing west. Repeat operations 
1 to 4 in this position, and take for the magnetic E-W the 
mean of the two directions so determined. 
If the edges of the frame are not strictly at right angles to the 
lines of force at its centre, but are inclined at an angle of ^ -f a to 
them, the two determinations will differ by 2a, and their mean will 
be independent of a, provided the edges of the frame are strictly 
parallel, and the magnetic declination is constant during the 
operation. 
Experiments with the above apparatus have shown that the error 
due to any small excentricity on the part of the magnetometer is 
inappreciable. The magnetometer was purposely placed 2 mm. 
away from the centre towards different quarters, but no sensible 
