oy j be, 
f 
May 29, 1873] 
water under similar circumstances of level. In this case the 
sea-level corresponds to what is called the Absolute Zero 
of temperature. [It is well to observe here that it is the 
potential energy of the water, not the quantity of water 
itself, which corresponds in this analogy to the quantity of 
heat. In this simple remark we have all that is necessary 
to correct Carnot’s reasoning in so far as it was rendered 
erroneous by his assumption of the materiality (and con- 
sequent indestructibility) of heat.] 2nd, Temperatures 
thus defined correspond, as Thomson and Joule have 
shown by elaborate experiments, very closely indeed with 
those given by the air-thermometer—the absolute zero 
being about 274° of the Centigrade scale below the freez- 
ing point of water. I have made this digression as I 
shall have frequently to use the word temperature, and I 
shall always employ it in the sense just explained. 
The subject of Thermo-electricity of course includes 
all electric effects depending on heat, but in this lecture 
I shall confine myself to the production by heat of cur- 
rents in a circuit of two metals. 
The transformation of heat into the energy of current 
electricity was first observed by Seebeck in 1820 or 1821. 
His paper on the subject (Berlin Ac., 1822-3, or Pogg. vi.) 
is particularly interesting, as he gives the whole history 
of his attempts to obtain a voltaic current from a circuit 
of two metals without a liquid, and the steps by which he 
was led to see that heat was the active agent in producing 
the currents he eventually obtained. In this paper See- 
beck gave the relative order of a great number of metals 
and alloys in the so-called thermo-electric series, and 
showed that several changes of order occurred among 
them as the temperature was gradually raised. 
+ In a note attached to this paper, Seebeck recognises 
that in this further discovery he was anticipated by Cum- 
ming (who seems, in fact, to have made an independent 
discovery of Thermo-electricity). Cumming showed that 
when wires of copper, gold, &c., were gradually heated 
with iron, the deflection rose to a maximum, then fell off, 
and was veversed at a red heat. 
[Seeheck’s original experiment and Cumming’s extension of 
it were exhibited. ] 
You see that, keeping one of the copper-iron junctions 
at the temperature of the room and gradually heating 
the other, I produce a current which increases in in- 
tensity more and more slowly till it reaches a maximum, 
then falls off faster and faster till at last it vanishes and 
thereafter sets in the ‘oposite direction. We are still far 
below the melting point of copper, yet further heating up 
to that point produces but little additional effect. The 
reason of this will be apparent from some facts to be 
described towards the end of the lecture. At the moment 
of maximum current the two metals are thermo-electri- 
cally Meutral to one another.—The temperature in the 
present case is about 280°C. 
Seebeck pointed out that bismuth and antimony (to 
the choice of which he had been led by a very curious set 
of arguments) were very far removed from one another in 
the series, and therefore gave large effects for small dif- 
ferences of temperature. This is still taken advantage of 
in the Thermo-electric Pile, which, when combined with a 
sufficiently delicate galvanometer, is even now by far the 
most delicate thermometer we possess. It has recently 
enabled astronomers to detect and measure the heat 
which reaches us from the moon, and even from the 
brighter fixed stars. In the skilful hands of Forbes and 
- Melloni this instrument was the effective agent in demon- 
strating the identity of thermal and luminous radiations— 
a step which, as regards the simplification of science, is 
as important as the discovery of magneto-electricity ; and 
which was completed by Forbes when he succeeded in 
polarising radiant heat. 
But when we come to look at this question from the 
point of view of transformation of energy, we have to ask 
NATURE 
be Ie 
87 
where is the absorption, and where the letting-down of 
heat, to which the development of the current considered 
as a rise of energy is due. Very remarkably, an experi- 
ment of Peltier supplies us with at least part of the 
answer. Peltier showed that, given a metallic junction 
which when heated would give a current in a certain 
direction, then provided a battery were interposed in that 
circuit (initially at a uniform temperature) so as to senda 
current in that direction, the passage of the current cooled 
the junction, while a reversal of the current heated it. 
This, considering the circumstances under which it was 
made, and the deductions since drawn from it, is one 
of the most extraordinary experimental discoveries ever 
made. Water was frozen, in an experiment by Lenz, by 
means of the Peltier effect. 
Here then is a reversible heat effect, and to it we may 
reasonably assume that the laws of thermodynamics may 
be applied ; although from the very nature of the experi- 
ment the reversible effect must always be accompa- 
nied by non-reversible ones, such as dissipation by heat- 
conduction, and by heat generated in consequence of the 
resistance of the circuit. The latter of these is in general 
smallin thermo-electric researches, but the former may 
have large values. 
It is known from the beautiful experiments of Magnus 
that no thermo-electric current can be produced by un- 
equal heating in a homogeneous circuit, whatever be the 
variations of section—a negative result of the highest 
importance. Sir W. Thomson, to whom we are indebted 
for the first and the most complete application of thermo- 
dynamics to our subject, showed that the existence of a 
neutral point necessitates the existence of some other 
reversible effect besides that of Peltier. And even if the 
circuit varied in section, the result of Magnus, just referred 
to, showed that this could only be of the nature of a 
convection of heat by the current between portions of the 
same metal at different temperatures. Thomson’s reason- 
ing is of the very simplest character, as follows :—Sup- 
pose the temperature of the hotter junction to be that of 
the neutral point, there is no absorption or evolution of 
heat there ; yet there is evolution of heat at the colder 
junction, and (by resistance) throughout the whole circuit. 
The energy which supplies this must be that of the heat 
in one or both of the separate metals; but reasoning of 
this kind, though it proves that there must be such an 
effect, leaves to be decided by direct experiment what is 
the nature and amount of this effect in each of the metals 
separately. By an elaborate series of ingenious experi- 
ments Thomson directly proved the existence of a current 
convection of heat, and (curiously enough) of opposite 
signs in the first two metals (iron and copper) which he 
examined. In his own words, “ Vitreous Electricity carries 
- heat with it in an unequally heated copper conductor, and 
Resinous Electricity carries heat with it in an unequally 
heated iron conductor.” This statement is not very easy to 
follow. It may perhaps be more intelligible in the form :—- 
In copper a current of positive electricity tends to 
equalise the temperature of the point it is passing at any 
instant with that of the point of the conductor which it 
has just left, z.c., when it passes from cold to hot it tends 
to cool the whole conductor ; when from hot to cold, to 
heat it, thus behaving like a real liquid in an irregularly 
heated tube. The effects in iron are the opposite ; and 
Thomson therefore speaks of the specific heat of electricity 
as being thus positive in copper and negative in iron. 
He givesavery remarkable analogy from the motion of water 
in an endless tube (with horizontal and vertical branches), 
produced by differences of density, due to differences of 
temperature. Here the maximum density of water plays 
a prominent part. Neumann has recently attempted, by 
means of the laws of motion of fluids, and the unequal 
expansibility of different metals, to give a physical ex- 
planation of thermo-electric currents. But, not to speak 
of the fact that positive electricity is by him considered 
