PROFESSOR THOMSON ON THE ELECTRO-DYNAxMIC QUALITIES OF METALS. 733 
Fig. 38. 
called the testing conductor, be soldered by its ends to the ends A and D of the con- 
ductor to be tested and of the reference conductor, or to strong pieces of metal to 
which those ends are firmly attached. Let one electrode of a galvanometer be 
soldered to the connecting arc BC, at its middle, or at any other point of it, Q; and 
let the other galvanometer electrode be ready to be applied by the hand to any posi- 
tion on the testing conductor. A current is then sent from 
one or more cells of Daniell’s battery through electrodes 
connected with A and D. This current flows through 
the divided channel ABCD and AP'OPD, in quantities in- 
versely proportional to the resistances of the two parts. 
The moveable galvanometer electrode is then applied, first 
to one point and then to another of the testing conductor 
(care being taken not to reverse, nor even to diminish, 
the magnetism of the lower needle in the astatic system of 
the galvanometer*), until by trial the point O that may 
be touched without producing any deflection in the needle, is found. The influence 
to be tested, whether it be magnetization, or tension, or elevation of temperature, 
is then applied to AB, or the influences to be tested against one another are applied 
to AB and CD, and the moveable galvanometer electrode is (if it has been removed) 
again applied at O. If the needle remains undisturbed, no effect is indicated ; that is, 
no alteration in the resistance of ABQ, or only an alteration in the same proportion 
as an alteration experienced by QCD, has been indicated. If, however, a deffection 
is observed, in such a direction that the moveable electrode must be moved to some 
point P in the part OD, it is inferred that the ratio of the resistance of ABQ to that 
of QCD has been increased ; or on the other hand, if such a deflection as requires a 
motion of the moveable electrode to a point P' in OA, the resistance of AB has been 
diminished relatively to that of CD. Fig. 39. 
150. As an example, I shall describe an experi- 
ment on the relative effects of tension on electric 
conductivity in copper and iron wires. Two pieces 
of stout copper wire. A, D, were each twisted into 
a loop which was made fast by solder ; a couple of 
inches towards one end of each wire being left free 
from the twisted part. These loops were put upon 
a strong hard wood peg about | of an inch dia- 
meter, at a distance of about J of an inch from one 
another; and to their lower ends were firmly 
soldered fine iron and copper wires (strong enough 
* In the galvanometers which I have used, the two needles of the astatic combination are of similar material 
(pieces of the same steel wire, tempered brittle), and the lower one is a little longer (perhaps by about 
than the upper. Both are magnetized to saturation, and consequently the lower preponderates and gives its 
5 n 2 
