4 REPORT ON THE COMPLETION OF 
the observatory, served to render the small room in which it was placed suffi- 
ciently comfortable for a computing-room. 
The Instruments.—The declination and bifilar magnetometers were made by 
Mr Gruss of Dublin. The magnets during 1841 and 1842 were enclosed by 
loosely fitting cylinders, resting upon the pillar tops, covered by glazed lids; but 
the junctions were neither pressed together by any contrivance, nor was there 
any moderate precaution taken to prevent currents of air from circulating in the 
boxes. In 1843 and 1844, these and various other defects were removed: the 
joints were lined with velvet and pressed firmly together; rectangular boxes 
were placed within the drums; both cylindrical and rectangular boxes were gilt 
outside and inside; and the whole apparatuses were covered by thick cotton 
hoods. The gilding prevented the effects of radiation in producing currents 
within the boxes; the other precautions prevented the entrance of external cur- 
rents, and secured to a considerable extent a uniform temperature and dryness. 
The suspension threads for the declination magnet broke at short intervals 
in 1841 and 1842; they were succeeded in June 1843 by a thread carefully pre- 
pared, which is still in use, and apparently as good as ever. In case, however, 
the suspension thread should have received any torsion, the magnet was removed 
occasionally, a brass bar was suspended, and the torsion was rendered zero for the 
magnetic meridian before replacing the magnet; when little torsion was found 
existing the trials were made seldomer, as the mere manipulation was apt to in- 
troduce torsion of itself. 
Errors due to twist in the suspension wires of the bifilar magnet, were 
eliminated by the process employed to determine the unit-coefficient. The 
interval of the wires was adjusted according to the instructions by the Committee 
of the Royal Society; but this rendered the instrument too sensitive, and, during 
moderate disturbances, the scale went out of the field of the telescope. This was 
attempted to be remedied at the time by turning the torsion circle till the scale 
reappeared, generally a difficult and unsatisfactory operation; in 1846. however, 
I employed a small magnet for the purpose, this being placed at known distances 
where its effect upon the bifilar magnet was determined. 
Even for the ordinary diurnal movements, however, the instrument was too 
sensitive; the reading of the magnet scale was estimated to one-tenth of a division, 
and of the thermometer to one-tenth of a degree, but a variation of one-tenth of a 
degree in the temperature of the magnet is equivalent to two-tenths of a scale divi- 
sion in its motion. The thermometer was placed with its bulb resting on a brass 
bar, so that it might show as nearly as possible the temperature of the magnet. 
The balance magnetometer was made by the late Mr Rosrson of London; 
the box was gilt inside and outside, and covered with cloth as in the case of the 
other instruments. The magnet was never lifted from its support, excepting 
