OF MAGNETIC DECLINATION. 353 
The angle y can be ascertained with the utmost accuracy by direct observa- 
tion, for it is the supplement of the zenith distance of the collimator-cross, as 
seen through the theodolite-telescope: but the ordinary methods of observation 
do not afford sufficient data for ascertaining the angles a and @. I have found, 
however, that these angles may be computed, provided we observe the magnet 
not only erect and ¢nverted, that is when turned round its axis 0° and 180°, but 
also when turned round 90° and 270°. 
8. In order to put this mode of observation in practice, I had a small colli- 
mating magnet constructed by Mr Apin, shown in fig. 3, consisting of a hollow 
steel cylinder 2-1 inches long, 0°5 inches in diameter, and about 0:04 inches thick, 
furnished at the end N with an achromatic lens of about 2 inches focal distance, 
and 0:3 inches aperture, and at C with a diaphragm carrying a cross of fine 
spider-lines. The diaphragm is supported by four screws, s, having their beads 
so deeply sunk below the surface of the cylinder as to be out of risk of dis- 
turbance; and the cell of the lens is made to screw into the cylinder, in order 
to adjust the cross-wires to its principal focus. The cylinder having been placed 
in temporary Ys, the lens was screwed in, until the wires were distinctly visible 
through the theodolite-telescope, which had previously been brought to sidereal 
focus; and then, by means of the screws s, the diaphragm was adjusted until 
its cross-wires continued nearly to intersect those of the theodolite-telescope, 
while the cylinder was turned round in the Ys. In this manner, the line of col- 
limation was rendered nearly parallel to the axis of the cylinder, and thus also 
approximately parallel to the magnetic axis. 
9. On the cylinder there is a tightly fitting brass ring A, having on its sur- 
face four lines, two of which, marked 1, 4, are shown in the figure; and the cy- 
_ linder itself turns without much friction in the ring B, to which the torsion- 
fibre of silk, F, by which it is suspended, is attached. The lines on the ring A 
are situated 90° from each other, and are marked for reference 1, 2, 3,4; and 
the ring B having also a line marked 0 upon it, with which the lines on A may 
_ be made successively to coincide, the cylinder may be suspended, either in its 
usual position, or turned through 90°, 180°, or 270°. The ring A was turned on 
_ the cylinder, until, when the lines 1 and 0 coincided, the wires of the collimator 
_ were sensibly horizontal and vertical,—a condition which was known to be ful- 
_ filled when, on looking through the theodolite-telescope, it was possible, by means 
_ of the tangent-screws, to make the vertical and horizontal wires of the theodolite 
_ cover those of the collimator. 
* In figures 4, 5, and 6, AB, CD are the wires of the theodolite-telescope, and ¢/, 
_ gh those of the magnet-collimator, represented, for the sake of simplicity, as they 
would appear when viewed through an erecting eye-piece. In figure 4, the mag- 
__ het is in its usual or erect position, which is known to'be the case when a small 
index 2, projecting from the edge of the diaphragm, is seen uppermost. In fig. 5, 
VOL. XXI. PART II. 5c 
