Prof. Norton on Molecular Physics. 277 



waves of translation, or in currents of free electricity, or both 

 combined, according to the conducting-power of the medium 

 exterior to the surface (p. 111). Substances may differ in the 

 effects thus produced under similar circumstances, from two- 

 causes — a difference in their conducting-power for heat, and a 

 difference in the degree of expansion, or in the effect of the ex- 

 pansion produced by the same amount of heat. 



Now let a plate of bismuth be placed in contact with a plate 

 of antimony, and let the junction be heated and the other ends 

 be brought into good conducting communication. If the above- 

 mentioned effects of heat be different from these two metals, a 

 current should set, at the junction, from the one which experi- 

 ences the greatest effect to that which experiences the least, and 

 pass through the circuit. Bismuth is a poorer conductor of heat 

 than antimony, and in fact than most other metals ; and hence 

 its surface-molecules should imbibe and retain more heat than 

 those of the antimony. The diamagnetic properties of bismuth 

 also indicate, as will be seen hereafter, that its molecular atmo- 

 spheres are remarkably expansible. If these peculiarities of bis- 

 muth be admitted, we have an explanation of the fact that bis- 

 muth is positive* to other metals in its thermo-electric relations. 

 The bismuth and antimony in the thermo-electric pair, it will be 

 observed, hold the same relation to each other as the zinc and 

 copper in the galvanic pair, and the heat does the same electrical 

 duty in the one that the oxygen does in the other. 



If the other ends of the two metals be brought together and 

 cooled, the current will be reinforced, since the molecular atmo- 

 spheres of the bismuth will contract more than those of the anti- 

 mony. (See effect of cold, p. 276.) 



It has been ascertained as the result of numerous experiments, 

 that "increasing the temperature of the negative metal gene- 

 rally increases the amount of deflection of the galvanometer- 

 needle produced by heating the junction ; while if the higher 

 heat is applied to the metal which is positive at moderate tem- 

 peratures, a current in the opposite direction is established." To 

 get at the explanation of these curious effects, we must observe 

 that the " higher heat " spoken of is applied at a certain distance 

 from the junction, and hence it is the inner sides of the surface- 

 molecules which first receive the heat from this second source, 

 and become positively polarized. Accordingly the current should 

 be strengthened in the first case above mentioned, and weak- 

 ened in the second. It has also been observed that " a current 



* By the positive metal is here meant that which imparts, at the heated 

 junction, positive electricity to the other metal. This is often termed the 

 negative metal. Its polarization is positive, but its electrical state is nega- 

 tive (p. 104). 



