The determination by experiment of the thermo-electric relations of any one substance belonging to the electromotive series to all other such substances is sufficient to fix all mutual thermo-electric relations among these. The first endeavour of the experimenter is then to obtain as convenient a substance for this purpose as possible. In investigating charcoal and certain alloys, we have in almost all cases employed one or other of two alloys of platinum and iridium, which have been already used by Professor Tait for a like purpose. The wires we used were the same which he discusses in his “First Approximation to a Thermo-Electric Diagram,” under the names of M and N. Their complete freedom from oxidation, their elasticity, and the high temperatures of their fusing points, rendered them peculiarly suitable for thermo-electric investigations through long ranges of temperature.

Generally both the M and N wires were firmly bound, each by its one extremity to the end or ends of the wire or wires respectively which were under investigation, in a multiple junction. This triple or, as it was in some cases, quadruple junction constituted the “hot junction.” The free extremities of the wires thus united were each bound to a moderately thin copper wire by very thin wire of the same metal, and the copper wires were led from these “cold junctions” to a commutator, which was in connection with a galvanometer. The commutator consisted of an arrangement of small mercury pools, into which the galvanometer and circuit wires, carefully amalgamated to ensure contact, dipped. All the junctions were formed in the manner indicated above, namely, by tightly binding the extremities of the wires together by thin copper wire.