Proceedings of the Cambridge Philosophical Society. 175 
A new proof of the propositions respecting the composition of 
forces, by J. King, M. A., Tutor of Queen’s College was read. 
April ^8 . — A paper was read by Professor Gumming, On 
the Relative Polarities of the Metals as developed by Heat, 
After noticing the apparent similarity between the well knomi 
experiment of exciting by heat the opposite states of electricity 
in the tourmaline, and the experiments of Dr Seebeck, on a bar 
of antimony, related in the last Number of this Journal, the 
magnetic effects of heat on other simple metals and alloys is ex- 
amined. All the metals were found to exhibit these effects, 
provided the extremities of the metallic bars were at different 
temperatures ; whether this difference were produced by heat- 
ing one extremity, or by cooling the other. 
Antimony and bismuth were most remarkable as producing 
greater deviations than any other metals ; and, under the same 
circumstances, the one being positive, the other negative. The 
electric states of other metals depended on the nature of the' con- 
necting wires, and seemed to be modified less by the galvanic 
relations of the metals, and connecting wires, than by their 
powers for conducting heat. For instance, the deviations caused 
by bars of copper, silver, or palladium, were positive or nega- 
tive, according as the connecting wires w^ere of silver or platina. 
Deviations were also produced, when the bars and wires were 
composed of the same metal. The electro-magnetic effects ap- 
peared not to depend upon any peculiarity of structure, such as 
crystallization, for they were not perceptibly altered by the man- 
ner of cooling the bars after fusion, and they were exhibited by 
fluid mercury inclosed in a glass tube. When a bar was broken 
and connected again by soldering, it acted as two distinct bars. 
The effects were not materially increased, by increasing the di- 
mensions of the metallic bars, or of the surfaces in contact ; but 
were much augmented by using wires of large diameter, until 
they reached ^^^th of an inch, after which there seemed to be no 
increase of power. When the circuit was increased in geome- 
trical progression, the effects appeared to decrease in arithmetical 
progression. A bar composed of two rods of antimony and bis- 
muth, heated at both extremities, gave a deviation of 36° to a 
compass needle inches long. A battery formed of 8 plates of 
antimony and bismutli alternately, gave some increase of power. 
