Lovering and Bond on Magnetic Observations at Cambridge. 73 
h, or more easily by inverting the Transit at B and using it as a colli- 
mator. If the Variation-transit be directed to the mid-wire of the 
large Transit, so that while its line of collimation falls behind the 
latter it also intersects the mark on Blue Hill, then we know that 
it is in the same meridian line as the Transit, and by turning it over 
to the North and reading off the azimuth of the line ae we find 
the variation of the fixed telescope and, consequently, the absolute 
value of every reading that is made with it. We pass now to the 
Horizontal Force Magnetometer which differs from the Bifilar Mag- 
netometer of Gauss only in employing a movable collimator instead 
of the mirror. The principle of either instrument is easily explained. 
A magnetized bar is suspended by a double wire, attached to two 
points near its centre at a definite distance apart, which remains 
the same through their whole length. The cap at the top from 
which the suspension is made is then turned until the torsion suffices 
to bring the bar round at right angles to the magnetic meridian. 
The horizontal component of the magnetic force of course tends to 
bring it back to the magnetic meridian, and therefore its mean value 
is measured by the mean torsion. But as the torsion remains near- 
ly constant at the same angle, any change in the horizontal force is 
indicated by the motion of the bar to the one side or the other of 
the perpendicular direction. This small variation from a mean po- 
sition is observed in the same way precisely as the variations of dec- 
lination in the first instrument; the change of horizontal intensity 
is a simple function of this small arc and very easily determined. 
When the absolute horizontal intensity is required, it is found by 
the method which Gauss employs with his Bifilar Magnetometer. c 
is the place of the Horizontal Force instrument. / is the place 
of the fixed telescope whose position at right angles to the mag- 
netic meridian is readily adjusted by the Variation-transit at a. 
10 
