9 4 D O C U M E N T 7 6 O N S U P E R C O N D U C T I V I T Y superconductor in which a current is flowing has no other motion than the one mak- ing up the electric current. Such a notion is made improbable not only by the Rutherford-Bohr theory, according to which there are strong electric fields inside a body, but also from the fact that superconduction is destroyed by moderate mag- netic fields. For, the lateral forces produced by the Lorentz force (Hall force)[15] would, of their own accord, balance out electrostatically by charge accumulation on the surfaces, such that no effect of the magnetic field on the electrons would be expected. It therefore appears that electric conduction must be attributed to the atoms’ peripheral electrons, which move around the nuclei at great speed. Indeed, accord- ing to Bohr’s theory, it scarcely seems conceivable that the energetically orbiting peripheral electrons would lose a considerable portion of their speed, i. e. in the case of mercury vapor with its relatively low-energy liquefaction. This is why, based on the current state of our knowledge, it looks as though free electrons do not exist at all inside metals. Then, metallic conduction would have to consist in atoms exchanging their peripheral electrons. If an atom were to receive an electron from a neighboring atom without roughly simultaneously releasing an electron of its own to another neighboring atom, it would experience a powerful energetic modi- fication, which certainly cannot happen with superconductive currents that main- tain themselves without any expenditure of energy. Thus it appears unavoidable that superconductive currents are borne by closed chains of molecules (conduction chains) whose electrons experience constant cyclical exchanges. That is why Kamerlingh Onnes compares the closed currents in superconductors to Ampère’s molecular currents.[16] With our broad ignorance about the quantum mechanics[17] of composite sys- tems, we are far away from being able to condense this vague idea into a theory. We can only attach a few questions to it that could be decided by experiment. It appears unlikely that different kinds of atoms could form conduction chains with one another. Thus the transition from one superconductive metal to another one may never be superconductive.[18] A further idea that suggests itself is that this may be the reason why thus far only metals with relatively low melting points have turned out to be superconductive [19] for, in such substances, the impurities may not be in a truly dissolved state but rather in the form of small complexes which, in the metal’s plastic state, will precipitate out. Furthermore, there is the possibility that the conduction chains can not carry arbitrarily small currents, but only those of a particular finite magnitude, which would likewise be accessible to experimental verification.[20] [p. 434]