DECLINOMETER. xv 
This value has not been used for this correction; it has been employed in the 
observations of absolute horizontal intensity made in 1845 and 1846. 
11. The second and most important error due to the torsion force is that pro- 
duced by the variation of the plane of detorsion. Unless when the period and extent 
of change is known, this can only be corrected by removing it; this is done occa- 
sionally in the following manner :—The magnet, with its stirrup, being removed, a 
brass bar and stirrup of nearly the same weight aud dimensions is suspended ; the 
rectangular box being removed, the cylindrical box being completely closed, and the 
shutters removed from the glazed lid, the extremities of the are of vibration are 
observed through the latter ; the marble slab beneath having radii drawn for every 
5° on each side of the magnetic meridian, the positions of rest at the extremities 
of the ares are estimated to 4°; the deviation of the mean position from the magnetic 
meridian is known, and the arms of the torsion circle are turned through an equal 
angle in the opposite direction ; the position of rest is then in the magnetic meri- 
dian. Much time and care were bestowed upon these observations, as the error 
due to this cause is by far the most serious that occurs in connection with the de- 
clinometer. 
12. The following are all the observations for the elimination of torsion made 
in 1845 and 1846, together with all the occasions on which the magnet was touched 
or removed from its box. When the mean position of rest for the north end of the 
brass bar was found to the east of the magnetic north, the torsion existing is con- 
sidered positive, and the effect of + 10° of torsion=—084 (from the previous obser- 
vation for the value of ©.) In the first case below, the north end of the brass bar 
was found to rest 7}° west of the magnetic north ; and the brass bar was thereafter 
made to coincide with the magnetic meridian. 
Feb. 54 34" 1845. Torsion removed, — 74°. The torsion could not have been 
more than 3° at first, but the north end of the brass bar moved slowly westward as 
if affected by currents, or as if the suspension thread were affected by moisture. 
March 234 232" 1845. Torsion removed, — 183°. 
April 24 2%", 1845. Torsion removed, — 34°. The line of detorsion of the 
supension thread varied during the observation; at first it was about N. 3° E. and 
8. 3° W.; ultimately it was N.5° E. and S. 5° W. The torsion circle was turned 3}. 
May 84 33", 1845. Torsion removed, — %. June 184 21" 1845. Torsion re- 
moved, + 14°, 
Sept. 214 22)", 1845. Torsion removed, + 19°. This amount of torsion may 
possibly be due to the dampness arising from washing the floor of the Observatory. 
September 204, the brass bar was suspended, with some difficulty in the manipula- 
tion, but it is not conceived that much torsion could have been thus introduced. 
Dec. 297 03", 1845. Torsion removed, 0°. The magnet with the short scale 
was placed in the declinometer box after this examination of torsion, the long scale 
magnet being removed to the intensity house for an observation of absolute hori- 
