NOTES REGARDING THE DECLINOMETER. 1x 
yary, so that the removal of torsion at any time does not guarantee that the 
needle shall always henceforth remain without it. The method adopted to pre- 
vent any considerable error arising from this source, is to eliminate the torsion at 
intervals, the declination needle being frequently compared with another, both 
before and after the elimination, in order to find the effect produced upon the 
reading of the former by eliminating the torsion. These comparisons will be found 
below among other notes regarding the declinometer. 
No correction deduced from them has, however, been made upon the daily 
observations of the declinometer, since the amount of error on account of torsion 
is always exceedingly small. 
NOTES REGARDING THE DECLINOMETER. 
1847, May 12¢ 5". Torsion circle turned from B 2213° to 236°. A comparison 
made between the declinometer and another magnet, before and after the elimina- 
tion, gave the effect of torsion before change = + 1/32. 
May 22° 6". Torsion circle turned from B 236° to 233}°. A similar com- 
parison gave the effect of torsion before change = — 042. 
Sept. 34.0". Torsion examined, circle turned from B 2334° to 242°. A com- 
parison between the declinometer and another magnet, before and after the elimi- 
nation, gave the effect of torsion = + 166. 
1848, Jan. 64 22". Torsion removed, circle turned from B 242° to 239°. A 
comparison as before seemed to show that the effect of torsion before the change 
was = + 0°06. Assuming the true effect to have been = — 025 (see Observa- 
tions for 1845-6, page 298, note), there is left an irregularity = 031. 
| From January 6th to January 18th, 1848, the needle in the vertical force 
magnetometer was sometimes in its place, and sometimes away ; and on January 
18th it was finally arranged in a plane parallel to the magnetic meridian, having 
previously been in a plane perpendicular to that meridian. Owing to this change 
of plane, the effect of the balance needle on the declination needle after January 
18th, 1848, was different from what it had been previous to that date. 
With regard to this effect, the correction, applicable to the declinometer scale 
readings on account of the balance magnet, while it was in a plane perpendicular to 
the magnetic meridian had been determined carefully (Sept. 4, 1843, and Jan. 25, 
1844) to be = — 0°80 scale divisions (see Introduction, 1843, p. xvi.) . . . . 
After Jan. 18¢ 5", this correction was + 0°60 scale divisions, as determined Jan. 
24, 1848, so that the old tables being used into which the correction — 0:80 
entered, a correction of + 1:40 scale divisions has been applied to the readings of 
the declinometer for every observation after Jan. 1845". The observations between 
Jan. 64 22" and Jan. 184 5” have likewise been suitably corrected. 
On Sept. 4° 23", steps, containing iron, were placed near the declinometer, 
from which they were removed Oct. 442. Jt was found that the eféct of the 
