
in Alloys of Bismuth and Tin. 561 
Lord Kelvin expressed the relation between the quantity 
of heat evolved or absorbed, the current and the temperature 
gradient in the form dQ=Cod@, where dQ is the amount 
Of heat evolved or absorbed by the current C in passing 
between two sections whose difference of temperature is dé@. 
The quantity o he called the specific heat of electricity. 
Comparative values of the specific heat of electricity for a 
considerable number of metals were obtained by Le Roux *, 
whilst absolute measurements have since been made by 
Haga f, Batellif and King §. In the experiments described 
in this paper the method used by Haga has been adopted. 
Briefly, the method is as follows :—The temperature-gradient 
is obtained by immersing the ends of the experimental rods 
in baths at definite temperatur es, and the value of the specific 
heat of electricity is found by comparing the change ot 
temperature at a point when a current flowing along the 
temperature-gradient is reversed with the rise in temperature 
produced at the same point by a current in the bar when this 
is at a uniform temperature throughout. 
In this latter case the heat generated is known from Joule’s 
law, so that, assuming that within the limits of the experi- 
ment change of temper ature is proportional to heat developed, 
the amount of heat produced or absorbed in the former case 
is at once calculated. 
The Thomson effect in bismuth has already been investi- 
gated by Batelli||, but for the sake of comparison, experiments 
were first made with bismuth and then with alloys containing 
increasing amounts of tin. 
Il. Preparation of the Specimens. 
The bismuth used in the experiments was obtained as pure 
as possible by procuring the pure oxide and reducing this in 
porcelain crucibles with pure potassium cyanide. In this 
way, bismuth containing no impurity other than 0°02 per 
cent. of iron was obtained. The tin used was supplied as 
pure precipitated metal. 
he specimens with which the experiments were made 
were cast in the form of rods about 35 cms. in length, and as 
thin as their brittle nature would allow—that is, about 5 mm. 
in diameter. In the case of the alloys the moulds in which 
the rods were cast consisted of hard glass tubes, of the 
* Le Roux, Ann. de Chimie et de Phys. x. p. 258 (1867). 
+ Haga, Ann. de Vécole polyt. de Delft, i. p. 145 (1885) ; 111. p. 45 (1886), 
t Batelli, Accad. delle Sci. di Torino, Atti, xxii. p. 548 (1887). 
§ King, Amer. Acad. Proc. xxxiii. p. 353 (1898). 
| Batelli, aii . 
Phil. Mag. 8. 6. Vol. 7. No. 41. May 1904 2Q 
