734 PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 
to bear weights of about 8 lbs. and 5 lbs. respectively). These wires were cut to the 
same length of 4^ feet, and their lower ends were put into slits about ^ of an inch 
deep, cut in the top of a piece of stout copper slip of the form and dimen- Fig. 40. 
sions shown in the diagram, and the copper pressed upon them to hold 
them fast by a pair of pincers. Solder was then applied, to make a com- 
plete and compact metallic connexion between the wires and the copper 
piece. A testing conductor, consisting of seven yards of No. 1 8 copper wire, 
was soldered by its ends to the upper copper pieces A, D ; and a current 
from six small cells of Daniell’s was sent through the double channel by 
electrodes soldered a little higher up to the same copper pieces. A, D. One 
galvanometer electrode was soldered to the lower copper piece, and the 
other was applied to the testing conductor till the point O, equipotential 
with the point of attachment of the former, was found. As from previous 
experiments I knew that an accidental variation of of an inch in the 
position of the moveable electrode on the testing conductor might lose or 
overbalance the effect looked for, I added a multi'plying branch, TFO'EU, 
consisting of a yard of No. 18 copper wire, with its ends soldered about 
half an inch on each side of O. This, of course, when touched by the 
moveable electrode, gave about thirty-six times the motion that would be 
required to produce or to correct any effect on the galvanometer if the 
simple testing conductor were used. The point O', on the multiplying 
branch, that could be touched without giving any deflection was then found ; and 
weights were hung from the lower end of the lower copper piece, so as to stretch the 
copper and iron wire equally. Immediately a deflection of the needle in the galva- 
nometer showed a current. This was corrected by sliding the moveable electrode 
on the multiplying branch towards U, that is, towards the parts conterminous with 
the copper wire. When the weights were removed, immediately a reverse deflection 
was observed. The conclusion is, that iron and copper wire equally extended have 
their resistances altered differently when under the stress ; that of the iron wire being 
more increased, should the absolute effect in each wire be an augmentation of resist- 
ance, as other experiments I have made give me reason to suppose it is, or less 
diminished should it turn out that the absolute effect in each wire is a diminution 
of resistance. 
151. Again, a heavier weight was applied so as permanently to elongate the wires. 
direction to the system. The strongest current through the coil only confirms the required state of magneti- 
zation, provided when it is started the index is either at zero, or on the side of zero towards which the deflec- 
tion is to he. If by accident a powerful current is admitted through the coil when the index is on the wrong 
side of zero, the lower needle has its magnetism instantaneously reversed ; but it may be as instantaneously 
put right again by suddenly reversing the current. If at any time, from the lower needle having either lost 
magnetic moment, or acquired a reverse magnetization, the astatic system is found reversed, it may be put in 
order with ease either by simply sending a powerful current through its coil, or by doing so and then suddenly 
reversing the current. 
