417 
of Edinburgh, Session 1874-75. 
Cumming experimented on many other pairs of metals, and, 
amongst them, copper and iron. In experimenting upon copper 
and iron, he found that as the hot junction is made hotter and 
hotter, the current increases, but only up to a certain limit. 
When the hot junction is made still hotter than a certain critical 
temperature, the current begins to diminish in strength, till it 
becomes zero. When the hot junction is made hotter still, the 
current becomes reversed. This was a very great discovery 
indeed in thermo-electricity. I could be tempted to go into 
some more details, but I should perhaps tax the patience of the 
Society were I to do so. But I cannot refrain from mentioning 
the thermo-dynamic relations of this discovery and the collateral 
discovery by Peltier, that there is a thermal effect produced at a 
junction of two different metals when by any means an electric 
current is caused to cross from one metal to another. I call 
that a collateral discovery, — collateral with Seebeck’s, because it 
might have been made quite independently of Seebeck’s. That 
might be one discovery and Seebeck’s another, thoroughly dis- 
connected, had we not light from theory to put them into relation 
with one another. The first ray of light thrown on the subject is 
to be found in an almost casual remark made by Joule in the 
course of some not casual — but thoroughly worked-out — observa- 
tions upon the relations between the generation of heat on the one 
hand and the development of work from heat on the other. He 
remarked that the absorption of heat or the production of cold, 
manifested under certain circumstances by electric currents cross- 
ing a junction of dissimilar metals in Peltier’s discovery, was to be 
looked to as a source of the power to be developed from a current 
of electricity produced by a thermo-electric action. The thorough 
working out of this remark of Joule’s required the application of 
Carnot’s theory. The relation between heat absorbed in one junc- 
tion of the thermo-electric circuit, heat evolved in the other, and 
work done by the current in any instrument driven by the current 
put into the circuit — let us say, a mechanical engine raising 
weights, driven by the current — the relation, I say between heat 
absorbed in the one junction, heat developed in the other, and 
mechanical work performed, was fully worked out by the applica- 
tion of Carnot’s theory. The introduction of Cumming’s inver- 
