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 9 3 and that the resistance of nonsuperconductive metals becomes temperature-inde- pendent at low temperatures. The curvature of the resistivity curve at low temper- atures is thereby indirectly connected to quantum theory. In fact, according to the above conception, as the temperature drops, the resis- tance of nonsuperconductive metals ought to approach zero, whereas in reality it approaches a limit other than zero. Kamerlingh Onnes found, however, that this limiting value is strongly influenced by slight impurities.[11] He found out, further- more, that these trace admixtures cause a vertical parallel shift of the entire resis- tivity curve, i.e., that they produce an “additive resistance,” so that the resistance of the pure, homogeneous metal may very well have the limiting resistance of zero.[12] Let it be said that this extremely remarkable fact obstinately resists explanation by equation (1). For, if the admixture creates special collision opportunities for the electrons, this causes, as is easily demonstrated, a constant contribution of . This latter does not change the resistance by a temperature-independent amount but by one proportional to u but u cannot by any means be assumed to be temperature-independent, because otherwise the sole great success of the theory, namely, the explanation of the Wiedemann-Franz law, would have to be aban- doned. For the same reason, too, it would be difficult to explain theoretically the resistance of impure metals becoming constant at low temperatures. From this sketch one can see that the thermal theory of electrons fails already for the usual conduction phenomena, let alone for superconductivity. On the other hand, it certainly is conceivable that the Wiedemann-Franz law will result from another kind of theory that relates electrical and thermal conductivity back to an electron mechanism. The failure of the theory became fully obvious after the discovery of supercon- ductivity in metals.[13] By proving that nonsuperconductive wires with a thin coat- ing of a superconductive material are superconductive, Kamerlingh Onnes convincingly demonstrated that superconductivity certainly cannot be based on electrons in motion from thermal agitation.[14] With time the coating’s electrons would have to penetrate into the nonsuperconductor and there lose their mean motion advantage that forms the electric current. Consequently, the system would not be superconductive. If one wished to explain superconduction by free electrons, one would have to conceive of them as agitation-free in such a way that the negative electricity in the [p. 432] 1 l -- - or to u n --, - resp. [p. 433]