PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 735 
The deflection, which was in the same direction as at first, was noted, but not cor- 
rected by any motion of the moveable electrode, and the weight was again removed. 
The needle returned towards zero, but remained deviating in the same direction as 
it had done to a greater degree with the weight on. By applying the hand instead 
of weights and gradually pulling down the lower copper piece, at first slowly, and 
afterwards rather faster, the needle could be made to deviate to 7° and kept steadily 
there. After the wires had been stretched by rather more than an inch, the hand was 
removed with a gradual diminution of stress, which could easily be regulated to let 
the needle down without oscillation to whatever position it would rest in, with the 
stress entirely off. This in several repetitions of the experiment on the same wires 
was found to be somewhere about 3° or 4° in the same direction as the deviation which 
was kept at 7° for a few seconds during the stress. Hence it was further concluded, 
that, as regards electric conductivity of the substance, the effect of permanent elonga- 
tion, remaining after the stress is removed, differed between iron and copper in the 
same way as the effect of longitudinal stress during its action ; that is, that the gal- 
vanic resistance of iron is more increased by permanent elongation than that of copper. 
Irregular variations to a considerable extent, obviously due to thermo-electric effects 
from the copper and iron in the compound conducting circuit, made me not attempt 
to measure with much care, the distance the moveable electrode had to be shifted to 
counteract the effects of tension; but I intend repeating the experiment and making 
it for other pairs of metals, with this source of irregularity removed by a modification 
of the testing conductor. 
152. In the kind of experiment which has been described, the channels through 
the two metals experienced exactly the same elongation, and, it may be said without 
committing any sensible error, the same narrowing, by the longitudinal extension. 
The effect observed, therefore, depends truly on variations in the conductivities of 
their substance. I had made previously various experiments on copper wire alone, 
and on iron wire alone, in which I attempted to eliminate the effects of elongation 
and narrowing, and had very nearly established, for the case of iron wire at least, 
that the augmented resistance due to tension, either temporary or permanent, is a very 
little more than can be accounted for by the change of form. As, however, I have 
other experiments in progress, by which I hope to be able to show for a single metal 
the absolute effect on its specific conductivity separated perfectly from any influence 
on the resistance of the conductor occasioned by a change of its form, I defer in the 
meantime giving more details of investigation on this subject. 
153. The method which has now been described has many great advantages over 
that by the differential galvanometer, or any other that I know of for testing or 
measuring galvanic resistances. In the first place, the irregularities, dependent on the 
electrodes, connexions, and circular conductors, of the differential galvanometer, are 
entirely done away with, and only the tested and the testing conductors, all con- 
nected by compact solderings, can influence the indication from which the results are 
