a 
50 Professor CUMMING on the Developement 
to itself, yet the deviations in the two last were in opposite direc- 
tions, though the contact of brass and platina was the same in 
both. 
If these experiments be referred to the hypothesis which ac- 
counts for electrical excitation by the oxidation of the metals, 
they seem equally adverse to it. Not to repeat the instances of 
its production where the heated bar and the wires were of the 
same metal, and in consequence, similarly, if at all, oxidated ; 
it can scarcely be imagined, that an elevation of temperature of 
not more than two or three degrees, should cause a difference 
of oxidation; and it is to be remarked, that the effect is produced, 
whether the temperature be elevated or depressed. On placing 
one end of a bar of bismuth in a freezing mixture, or even by 
allowing a few drops of ether to evaporate from its surface, there 
was produced a considerable deviation on the needle of the gal- 
vanoscope; that extremity which remained at the temperature of 
the room acting as the heated end of the bar in the other instances. 
Having ascertained that in all the perfect conductors of elec- 
tricity, electro-magnetism may be excited by the unequal distri- 
bution of heat, my next endeayour was to determine the direction 
in which this peculiar influence is exerted, and the mode of its 
propagation. If a bar of antimony, AB (Fig. 3.), having its ends 
connected by copper wire Aab B, be heated at one extremity 
and presented to the compass, the deviation, in every part both 
of the bar and the wire, is of the same nature, and therefore the 
current of electricity (if there be such a current) is throughout 
in the same direction. The effect is similar, whatever metal be 
employed; but, as will be seen by reference to Table I, the 
direction of the current in some is opposite to that in others. If 
two bars of the same or of similar* metals, equal in power, be 
* Similar, as to their developement of electro-magnetism by heat. 
