AND INSTRUMENTS AT MUNICH. 507 
oil ee 
sa 
nF WT 
The magnet M M with a strong wire A B affixed was suspended 
by the thread F’; one end of the spiral spring was attached to 
the strong wire A B at a, the other end was made fast at 6 to 
§ S, which projected from an insulated supporting pillar. The 
tension of the spring was such as to keep the bar 90° from the 
magnetic meridian. It will easily be conceived that by means 
of an attached mirror any alteration in the position of S S might 
be recognized, so that either the torsion force of the spring 
would be constant, or a suitable correction could be applied. 1 
will only notice further that I determined the effect of tempera- 
ture repeatedly and with great exactness; and as the tempera- 
ture of the bar underwent very little change, there have been 
few instruments employed, the results of which are deserving of 
equal confidence. 
In using the instruments several circumstances were speedily 
noticed, which appeared to require improved arrangements. 
a. The determination of torsion in the declination instrument 
is a very tedious operation, and at the same time one which 
the constantly occurring changes require to be performed very 
frequently. 
b. There is no check on the amount of the loss of magnetic 
force in the bar of the intensity magnetometer. 
ih 
scribed in the ‘ Reports of the British Association.’ [Also in Phil. ‘Trans. 1836. 
art. xx.—Ep.] Ingenious as is this arrangement, it can hardly afford in prac- 
tice all the advantages which it promises in theory. Silk threads are hygro- 
metric, and metal threads, according to ny experiments, have the disadvantage, 
_ when they extend or contract (from temperature or other causes), of at the 
same time becoming twisted. 
2n2 
