ane on Electric Conduction in Metals 239 
electrolytic conductors. On account of the peculiar relation of 
these bodies to me is impossible to apply to 
them the same form of experiment : I have used with the 
metals. I have however set on foot such modified forms of ex- 
periment as my means will allow, but cannot complete them in 
time for this paper. ‘Their success is doubtful on account of in- 
herent difficulties, but if successful they may be communicated 
on some future occasion.* 
17. A similar method of experiment to the above may be em- 
ployed for the exact determination of the conducting power of 
metals, as also the variation of their conducting power with va- 
riations of temperature, and the general law, if any, which gov- 
erns such variation, and other particulars. I have made a few 
experiments on conducting powers, but there was nothing of im- 
portance new in the results. A statement, however, made in this 
Journal} a few years since, that the conducting power of a wire 
was greatly impaired by bending or twisting it, was not confirm- 
ed. After winding and unwinding a thick wire several times 
over a cylinder less than an inch in diameter, the conducting 
power appeared scarcely affected, either in iron or brass wire.t So 
the annealing a hard iron wire at a low red heat affected its con- 
ducting power very little. 
18. In the preliminary experiment on the uniformity of con- 
duction in wire, the interference of thermo-electricity with the 
legitimate indications of the galvanometer was soon manifest. 
Such precautions however were taken, that in all the subsequent 
experiments no indications, or but doubtful ones, of thermo-elec- 
tric action were perceived when the battery was not in action. 
But when the battery current was flowing, the indications of the 
galvanometer were subject to fluctuations, as if the points of equal 
tension were not quite stationary. These fluctuations, though 
confined within very narrow limits, and by no means sufficient 
to affect materially the results, were still manifest, and were as 
great near the soldered extremities of the wires as in the middle 
parts. ‘They may therefore be referred to variable thermo-elec- 
tric action between the wires and solder, in consequence of heat 
* One of these experiments has been made since this was written, and the law 
found to hold in the case of dilute acid. 
See this Journal, first series, vol. xxxv, p. 109. 
t A copper wire gave the same result, but it had been previously bent. 
