722 PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 
131. The wires used, with the exception of the iron, steel and brass, were all sup- 
plied by Messrs. Matthey and Johnson, as chemically pure. The results of the 
experiments (made as described in §§ 120 and 121) ontbe effects of lateral hammering 
were, in every other kind of wire tried, the reverse of those found for iron. Thus in 
steel, copper, tin, brass, lead, cadmium, platinum, zinc, the current was always found 
to be from the unhammered to the hammered portions through hot. All the wires 
except zinc were carefully annealed by myself, before they were coiled and hammered 
(§ 120); but the process of annealing by heating in oil and cooling slowly made the 
zinc very brittle and crystalline, instead of softening it as in the other cases, and it 
was therefore taken as supplied by the manufacturers, and coiled on the core and 
hammered in the manner described. 
132. The experiments on the coils differently tempered in their different parts 
(§ 126), in the cases of tin and cadmium, gave only doubtful galvanometer indica- 
tions ; zinc wire proved so brittle in the annealed parts as to defeat some attempts to 
test the thermo-electric effects of temper. I have little doubt but that results may 
be obtained in all these cases by a careful repetition of the experiments, with perhaps 
some modification to meet the peculiarity of zinc. Slips of sheet iron and of sheet 
copper were tried without any thermo-electric indication being noticed. [Addition, 
Dec. 1856. — I have recently found in slips of sheet iron the same thermo-electric 
effect of temper as in round and flattened iron wires.] All the other conductors tried 
gave very decided results. In the cases of round iron wires of very different dia- 
meters, of iron wire flattened through its whole length by hammering, of round 
steel wire, and of steel wire flattened through its whole length by hammering, and of 
steel watch-spring, the thermo-electric effect of annealing portions of the coil after 
the whole had been suddenly cooled, was a current from unannealed to annealed 
through hot. In round wires of copper and brass, the thermo-electric effect of the 
same process was a current from annealed to unannealed through hot. 
133. The effects of permanent torsion were decisively tested only for iron and 
copper wires ; and they proved to be in each case the same as the effects of hardening 
by longitudinal extension, by lateral compression, or by rapid cooling, being quite 
decidedly yrom brittle to soft through hot in the iron, and from soft to brittle through 
hot in the copper. 
134. The views explained above (§ 105), by which I was led to look for the thermo- 
electric qualities of a crystal in a non-crystalline metal subjected to mechanical strain, 
show the probability of finding sucb properties also developed along with magnetism, 
by external magnetic force, especially in the few metals, iron, nickel and cobalt, 
which have high capacities for magnetic induction. Towards verifying this idea I 
tried first the following simple experiment, analogous to the first experiment (§ 107) 
which I had made on the thermo-electric effects of tension. A little helix about 
3 inches long, consisting of 220 turns of thin covered copper wire laid on in three 
strands on a cylindrical coie of pasteboard, about \ of an inch internal diameter, 
