Abstract
Conventional models of electronic conduction through localised states lead to a simply activated temperature dependence, log sigma varies as T-1, for the DC conductivity sigma . But this behaviour is not always followed over the whole experimental range, and in some instances log sigma varies as T appears to be a better description. The authors show some examples of this behaviour, and point out that it is a characteristic that was anticipated more than 20 years ago by a model of quantum tunnelling through a vibrating potential barrier. They show how this model can be incorporated into the small-polaron picture of electron transport to explain qualitatively a wider range of experimental behaviours.