The linear transport theory prescribes that in a semiconducting oxide, an ionic flow could be caused not only by the ionic electrochemical potential gradient, but also, indirectly, by the electronic electrochemical potential gradient. It has long been believed, however, that the latter - indirect or interference effect may be negligible. Quite recently, we have experimentally found for Co1-deltaO that the interference effect is by no means negligible. The results which have been accumulated up to date are reported, and the implications of the non-zero interference effect are discussed.