The success of the ionic hypothesis as an interpretation of the conductivity of electrolytes and gases has suggested the desire to try if a similar hypothesis can represent the ordinary conductivity of metals.  We are thus led to conceptions which at first sight seem audacious because they are contrary to our habits of mind.  They must not, however, be rejected on that account.  Electrolytic dissociation at first certainly appeared at least as strange; yet it has ended by forcing itself upon us, and we could, at the present day, hardly dispense with the image it presents to us.

The idea that the conductivity of metals is not essentially different from that of electrolytic liquids or gases, in the sense that the passage of the current is connected with the transport of small electrified particles, is already of old date.  It was enunciated by W. Weber, and afterwards developed by Giese, but has only obtained its true scope through the effect of recent discoveries.  It was the researches of Riecke, later, of Drude, and, above all, those of J.J.  Thomson, which have allowed it to assume an acceptable form.  All these attempts are connected however with the general theory of Lorentz, which we will examine later.

It will be admitted that metallic atoms can, like the saline molecule in a solution, partially dissociate themselves.  Electrons, very much smaller than atoms, can move through the structure, considerable to them, which is constituted by the atom from which they have just been detached.  They may be compared to the molecules of a gas which is enclosed in a porous body.  In ordinary conditions, notwithstanding the great speed with which they are animated, they are unable to travel long distances, because they quickly find their road barred by a material atom.  They have to undergo innumerable impacts, which throw them first in one direction and then in another.  The passage of a current is a sort of flow of these electrons in a determined direction.  This electric flow brings, however, no modification to the material medium traversed, since every electron which disappears at any point is replaced by another which appears at once, and in all metals the electrons are identical.

This hypothesis leads us to anticipate certain facts which experience confirms.  Thus J.J.  Thomson shows that if, in certain conditions, a conductor is placed in a magnetic field, the ions have to describe an epicycloid, and their journey is thus lengthened, while the electric resistance must increase.  If the field is in the direction of the displacement, they describe helices round the lines of force and the resistance is again augmented, but in different proportions.  Various experimenters have noted phenomena of this kind in different substances.

For a long time it has been noticed that a relation exists between the calorific and the electric conductivity; the relation of these two conductivities is sensibly the same for all metals.  The modern theory tends to show simply that it must indeed be so.  Calorific conductivity is due, in fact, to an exchange of electrons between the hot and the cold regions, the heated electrons having the greater velocity, and consequently the more considerable energy.  The calorific exchanges then obey laws similar to those which govern electric exchanges; and calculation even leads to the exact values which the measurements have given.[31]