550 
Proceedings of the Royal Society 
to the original and inverted positions of the coil, are to be read on 
the azimuth circle. 
To determine the astronomical meridian, the coil and the magneto- 
meter are to be removed without displacing the base, and the tele- 
scope mounted. Any of the usual methods of observation can then 
he applied. 
An alternative construction is to attach directly to the telescope 
a frame for the coil. This requires that the tube of the telescope 
be perforated in its middle to let in the magnetometer. The in- 
strument becomes somewhat clumsy, but it has the advantage of 
allowing the same pivots to he used in both the magnetic and 
astronomical observations. 
The sensibility of the apparatus (in either its simple or more 
elaborate form) may be investigated as fol- 
lows : — 
Let H he the horizontal component of the 
terrestrial field. 
Let F be the field due to the coil alone. 
Let 6 he the angle between F and H. 
Let 8 be the angle hetween H and E,, the 
resultant of H and F. This will he 
the angle through which the magneto- 
meter is deflected when the current is 
made, and 
± F sin ^ 
tan 8 = 
F cos 0 -f H 
F having a -f or - sign, according as its direction is the same or 
opposite to that of H. 
Since 8 and 0 are small, 
±F^ e 
±F4-H“-, h 
from which we see that 8, the deflection of the mirror due to any 
assigned error in the position of the coil, can be greatly magnified 
h}^ making F a little less than and of opposite sign to H. 
This magnification can he roughly measured by turning the frame 
through a known angle from its determined position, and observing 
