on the Electrical Resistance of Bismuth. 193 



because there is no certainty that the melted metal is always 

 cooled under the same conditions. The bismuth rod being 

 thus prepared, the two side branches which are not wanted 

 are taken away. 



The electrodes to convey the electric current remain to be 

 fixed. For this purpose one end of the capillary rod of 

 bismuth is heated at the same time as a platinum wire in a 

 Bunsen burner up to the melting-point of bismuth ; then the 

 red-hot platinum wire is plunged into the capillary tube. It 

 is easy to see that a good contact is thus produced. 



2. In order to obtain bismuth rods very rapidly cooled or 

 tempered, the melted metal was poured rapidly into a kind of 

 iron trough, cold, and forming a somewhat acute dihedral 

 angle. If the pouring takes place rapidly, by inclining the 

 trough sufficiently to cause the melted bismuth to run down 

 very quickly, rods of bismuth can be obtained sufficiently long 

 and not too thick. This last condition is indispensable, other- 

 wise the bismuth rods will have too small a resistance. 



No adherence takes place during this tempering between 

 the iron and the bismuth. Moreover, by taking certain 

 precautions, the surface of the bismuth may also be filed, and 

 a fresh part of the metal be thus exposed. 



In this case the electrodes were made of copper wire, 

 soldered to the two ends of the rod of bismuth by means of a 

 fusible alloy. 



3. The compressed bismuth which we used in our experi- 

 ments was given to us by Prof. Spring. Having been pre- 

 pared in a drawplate*, it was absolutely cylindrical. 



The electrodes were formed, as in the case of the tempered 

 bismuth, of copper wires soldered to the bismuth with a 

 fusible alloy. 



The molecular arrangement in this compressed bismuth is 

 very remarkable. A fracture normal to the axis of the 

 cylinder presents a radial structure. We shall see later on 

 the curious results that this sample has given us. 



The alloys have only been studied under two states : — 



(1) Melted and very slowly cooled, 



(2) Tempered. 



Rods of these alloys have been prepared in the same way 

 as those of bismuth. 



* Annales de la SocieU Geologique cle Belgique, t. xi. p. cxxxiv (1884). 



