716 PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 
the lower double iron beam, which is held down by the links and the bar under 
the roof of the press. Before working the press, 
the rectangular wooden frame with its iron cross- 
head is steadied by cords from hooks in the ceil- 
ing, and the following arrangements are made : — 
Two slips of sheet iron, each about 18 inches long, 
are soldered to the upper and lower ends of the 
oblique conducting channel, and their other ends 
are soldered to copper wires and put into the 
circuit of a galvanometer, with the usual precau- 
tions (§ 92) to ensure equality of temperature and 
electrical insulation between the two junctions of 
the dissimilar metals. Four tin-plate tubes, of 
semicircular section, each about f-inch diameter, 
and coated with a single fold of paper pasted 
round it, are pressed with their flat sides on the 
two sides of the sheet iron against the upper and 
lower edges of the oblique conducting band ; and 
are connected by india-rubber junctions, so that 
steam may be blown through two of them to heat 
one edge of the conducting band, and cold water sent through the other pair to keep 
the other edge of the band cold. The arrangements being thus made, a small boiler, 
heated by a common wire-gauze gas-lamp, is used to send steam through one pair of 
the tubes, and the town-supply water-pipes give a continued stream of cold water 
through the other pair. When the galvanometer was observed, there was at first no 
sensible indication of a current. The press was then worked, and the galvanometer 
immediately exhibited a slight deflection. The press was released, and a careful 
observation gave again little or no evidence of a current. Then, by an arrangement 
of double-branched stop-cocks, the steam and cold water were quickly reversed, so 
that the edge of the conducting band which was hot became cooled, and the other 
one became heated. Still the galvanometer showed no sign of current until the 
press was worked, when a reverse deflection to the former was manifested. While 
the press was kept up the steam and cold water were again sent along the same 
edges as at first. After a short time the deflection of the needle was reversed, and 
the same current as at first was indicated. The deflections were very slight in each 
case, but were unmistakeably demonstrated by the use of the reversing break (com- 
mutator) connected with the galvanometer. Had it not been for the accident noted 
above, a much more powerful stress would have been applied to the iron, and I have 
no doubt but that conspicuous deflections of the needle would have been produced. 
117 . The current in every case was down the inclined channel of sheet iron when the 
upper edge was heated, and up the incline when the lower edge was heated. That is, if 
Fig. 25. 
