702 PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 
an inch diameter brazed to it by ends of sheet copper, shaped as shown in the 
diagram. The space round the inner tube and within pig. 14 . 
the outer sheet and ends was filled with oil completely 
covering the inner tube, and, when heated, rising into 
the space between the upper parallel plane parts of the 
outer sheet. A narrow ring of sheet metal with a long 
slip projecting from one side for holding it by, was put 
in the inner tube before the other parts were brazed on, and during an experiment 
was kept as constantly moving from one end of the bath to the other and back as 
was required to keep the whole mass of oil at one temperature. 
The second drawing represents, on the actual scale, a section of 
either bath through the position occupied by its stirrer. This 
diagram also shows a section of an external case of sheet metal 
which supports the bath, and serves as a flue to carry the flame 
and products of combustion round its sides. The rows of gas- 
burners for the two baths were fixed in a line, and each burner 
regulated by a separate stopcock. The outer cases are screwed 
to the same stand, and the copper vessels holding the oil are 
pushed into them and rest with their axes in a line over the burners. 
The ends of the baths and of the outer cases are kept about one-fourth of an inch 
apart, and their supports are also made quite separate, whieh was found to be neces- 
sary to allow one of the baths to be kept cool, while the other is raised to a high 
temperature. (In the third drawing, the stems by which the stirrers are held are 
accidentally omitted.) 
89. When the baths and their furnaces are all fixed in their proper positions, a 
tube of thin glass, about 10^ inches long, just small enough to enter easily, is pushed 
Fig. 16 . 
into the inner tubes of the baths, and is left resting there, with its ends projecting a 
little outside their remote ends. In recent experiments I have substituted a simple 
roll of paper for the glass tube, and have found it answer quite as well. 
90. A compound conductor, to be tested thermo-electrically in this apparatus, con- 
Fig. 15. 
