PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 737 
tubes being 2 ,s thin and as well smoothed as it could be got. The piece of iron wire 
to be tested was soldered at one end to a piece of thick copper wire, and then insu- 
lated by a thin coating of writing-paper, wrapped twice round it, and pushed into 
the inner brass tube, which was just large enough to admit it easily. A second iron 
wire of equal dimensions was similarly prepared and inserted in a second core, in all 
respects like the other, except that in this experiment it had no copper wire wrapped 
round it. The two cores being laid side by side, the free ends of the iron wires were 
connected as shown in the diagram, by an arc of thick copper wire, C, soldered to 
them. A current from a single large cell of Daniell’s was admitted and carried off by 
the electrodes A and B. Cold water was kept constantly flowing through the spaces 
between the concentric brass tubes round the iron wires. The testing conductor 
(§149) used in this experiment consisted principally of the following parts : — (1) Two 
pieces of No. 1 8 copper wire, each sixteen yards long, prevented from touching one 
another by a piece of twine between them, rolled together on a thin copper cylinder, 
12 inches long and 3 inches diameter, from which they were insulated by a coating 
of two folds of silk cloth sewed round it. (2) Soldered to two of their contiguous 
ends, a connecting arc of thick copper wire, which was at first intended to be gradu- 
Fig. 41. 
ated, and will be called the scale of the testing conductor. (3) Separate short thick 
wires soldered to the other ends of the wires coiled on the copper cylinder, to bear 
binding screws for making connexions with the electrodes A and B of the conductor 
to be tested. One electrode of the galvanometer was soldered to the middle of the 
connecting arc between the two iron wires, and the other was held in the hand, and 
applied about the middle of the scale of the testing conductor. A rather troublesome 
process was then required to bring the galvanometer to zero by adding resistance on 
one side or the other between the ends of the testing conductor and A or B. When 
this was done, it was found that great deviations of the galvanometer needle were 
produced by sliding its moveable electrode a few inches in either way on the scale, 
and a perfectly sensible deflection by sliding it as much as |^th of an inch. The 
point of the testing scale to which the moveable electrode had to be brought, to give 
