PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 703 
sists of a wire or a thin bar of one metal, or a bundle of wires of two different metals 
5^ inches long, with wires of from 18 to 30 inches long of another metal soldered to 
its ends. To avoid circumlocution I shall call the former the mean conductor, and 
the wires soldered to it the electrodes, of the thermo-electric arrangement. The 
connexions between the mean and the electrodes are generally made by brazing, or 
by hard silver solder, when temperatures much above the boiling-point of water are 
to be used. 
91. A conductor thus prepared of two metals to be tested is drawn through the 
glass tube till the mean occupies a position, lying on the glass or paper tube, with its 
centre under the centre of the tube, and consequently with its ends about the middle 
of the hollow spaces surrounded by the oil-baths. 
92. The electrodes are carried from the ends of the insulating tube to the con- 
nexions required for completing the circuit through the coil of a galvanometer. These 
must essentially be maintained at the same temperature, unless the electrodes of the 
thermo-electric arrangement be copper, the same as those of the galvanometer. After 
trying several obvious, more or less troublesome plans to secure the fulfilment of 
this condition, I found a perfectly effective way simply to tie the connexions firmly 
together as close to one another as possible, only separated from contact by a fold 
or two of paper wrapped round each, and to tie a quantity of paper, or to make 
up a bundle of cotton wool, or some other bad conductor, round the two, for two 
or three inches on each side of the junctions. The junctions themselves, except 
when they are between homogeneous metals, are not made by binding-screws, but 
either by soldering, or by cleaning the surfaces and then tying the metal firmly 
together by fine twine. To avoid mistakes and prevent the necessity of disturbing 
the bundle round the junctions, in tracing the courses of the conductor on the two 
sides of it, a thread or mark of some kind is attached to one galvanometer electrode, 
and a corresponding mark on the electrode of the thermo-electric apparatus to which 
it is joined. This system of electric insulation and thermal connexion between 
junctions of dissimilar metals, I have found very convenient in a great variety of 
thermo-electric and other electro-dynamic experiments, and when it was used I have 
never observed the slightest trace of a current attributable to any difference of tem- 
peratures in the parts of the circuit to which it is applied. 
93. The conductor being thus arranged, two thermometers are pushed into the 
glass or paper tube from its ends and placed with the centres of their bulbs as close 
as possible to the metallic junctions, and with their graduated tubes extending nearly 
horizontally outside the apparatus, but inclined upwards as much as the inner dia- 
meter of the insulating tube and their dimensions permit, so as to check as much as 
possible the tendency (in some of the thermometers found very inconvenient) of the 
column of mercury to divide when sinking rapidly. All the space inside the glass or 
paper tube left vacant by the thermometers and the conductor is filled with cotton 
wool, well pressed in to prevent currents of air. 
94. This apparatus has many advantages over that first used and described in 
