PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 721 
consisting of round wires, of wires flattened by hammering, and of flat slips, of one 
metallic substance. 
127. Twisting a wire beyond its limits of elasticity hardens it perhaps as much as 
traction or hammering, and certainly in every case, when continued far enough, makes 
the metal very brittle. The nature of the mechanical strain here operative is easily 
expressed and explained in the theory of elasticity in terms of simple strains different 
in magnitude and direetion in different parts of the wire ; but it is not very easy to 
judge by theory from the effects of simple strains supposed known, what kind of 
thermo-electric effect, if any, is to be expected in a metallie wire, with strain thus 
heterogeneously distributed through it. I have therefore made experiments to deter- 
mine this effect in various metals. 
128. For experimenting on the thermo-electric differences between annealed and 
unannealed metallic conductors, a wire, round or flattened, or a slip of the metal was 
wrapped in a coil of from ten to thirty turns on a wooden core, about 2 inches broad 
and ^ of an inch thick, or sometimes only an inch broad, with a flat slip of thin 
sheet-iron laid on one side of it. The wooden core was then drawn away, and the 
coil, held in form by the thin iron core, was heated to redness in the fire, or to some 
temperature short of its melting-point, in hot oil, and was then suddenly plunged in 
cold water. After that, one side of the iron core was held over a flame, so as to heat 
the parts of the coil next it, while the parts of the coil on the other side were care- 
fully kept cool, by the constant application of cold water with a sponge. The 
wooden core was then slipped in and the sheet-iron removed ; and the coil was ready 
for testing by the galvanometer. 
129. The preparations for an experiment on the thermo-electric effect of permanent 
torsion, were commenced by bending a short portion at each end of a length of two 
or three yards of the wire to be examined, holding these end portions so as to keep 
the wire between them firmly stretched, and twisting it till it became brittle. It was 
then wound on a flat iron core (unless it was too brittle, as often proved to be the 
case, and then another wire was similarly prepared but not twisted quite so much) ; 
the parts of the coil on one side were carefully annealed by flame or hot oil, while 
those on the other side were kept cool by sponging with cold water. The iron core 
was then drawn out and the wooden core slipped into its place ; and the coil was 
ready for testing by the galvanometer. 
130. In making the thermo-electric experiments on the Fig. 29 . 
coils prepared in these various ways, glass heaters were 
first used, but I afterwards substituted two tubes of horse- 
shoe section made of tin-plate and coated with paper, which 
were applied with their concave parts touching the coil 
round its two edges. Steam from the small boiler was sent 
through one of these, and cold water from the town supply- 
pipes through the other. 
