towards the Theory of Thermo-electricity, 267 



VI. Can the currents which pervade a mass of bismuth [and 

 other metal), when heated at a point, be drawn off by con- 

 ductors, so as to be traced up and connected with its thermo- 

 negative property ? 



26. Mr. Sturgeon found that a solid prism of bismuth, heated 

 at a point near the end, gave out an electric stream, which 

 flowed along the side where the heated point lay, ran over the 

 edges before it reached the other end, and returned along the 

 opposite side to the point of heat*. 



I cast several slabs, prisms, and cylinders of this metal of 

 various dimensions, some being cast flat, others on end, but 

 could not obtain the regularity of circuit described by Stur- 

 geon. In every case in which a current was detected from the 

 heated point, another was found returning towards it ; but it 

 did not always happen that this was on the opposite side. 

 Sometimes, on a slab 4 inches wide, it was on the lateral parts 

 of the same face, more often on the edges : in several instances 

 the current was towards the heated point, along the middle of 

 the same face, and back again along the lateral parts or edges. 

 In many cases the current took the same direction on the op- 

 posite sides, and returned along the edges. One slab, cast 

 just at the melting point, in a cold mould, crystallized in gra- 

 nules; and in this no currents were detected. 



Each piece was tried with a point of heat about the middle 

 of each end of every face, allowing it to cool perfectly between 

 the successive heatings. The heat was applied by the dart of 

 the spirit blow-pipe, and at various degrees, a melted spot 

 being produced in a few seconds. The bismuth was instantly 

 raised close under the lower of a pair of short needles, not 

 quite astatic, the needles being 2 inches apart. They were 

 suspended by a fibre of unspun silk, within a glass shade, 

 closed at the bottom with writing-paper, on which was a di- 

 vided circle and a guide-cross as described (1.) p. 207, only the 

 cross was here drawn on the paper, as the thickness of card be- 

 tween the bismuth and the needle would have been too great in 

 some cases. 



27. A cylindrical bar (10 x f inches) was filed up to f. bright 

 surface, and two copper wires made to slide upon it, the ends 

 being pressed against it by a couple of card hoops, which slid 

 with the wires. The other ends of the wires were amalga- 

 mated, and dipped into the mercury boxes of the magnetest. 

 The bar being now heated at a given point, the direction of 



* Phil. Mag. and Annals, N.S. vol. x. p. 1 5. I attribute this to Mr. Stur- 

 geon, not having seen the description of Chevalier Yelin's experiments, 

 and therefore being unable to say whether he anticipated this one or not. 



2 M 2 



