Resistivity Changes: Another hypothesis would be that the resistivity itself changes with temperature. In order to understand why, we need a model of resistivity; one often-used simple model is the Drude model, 3 wherein the resistivity can be calculated as rho=m/n(q)^2τ
. Clearly, neither the mass m nor charge q of the electron will depend on temperature. Thus, for the resistivity to vary with temperature, either the number density of "conduction electrons," n, or the time between electron collisions with ions, τ, will change with temperature. What is a plausible reason that one of these factors might change with temperature? Does the direction of that change result in a change in resistivity that matches what you expect? (I.e.: would the temperature-dependent effect you note cause the resistance to increase or decrease? Does that match what you expect?)