Ion-selective micro-electrodes have been used to measure K+ and Ca2+ activity changes in extracellular space beneath the surface of the neocortex and cerebellar cortex during current flow across the tissue surface in anaesthetized rats. Inward currents produced decreases of [K+]o and outward currents produced increases, with insignificant changes in [Ca2+]o. Changes of [K+]o were largest just under the surface of the tissue, but were detectable down to depths of ca. 1 mm. With appropriate sitting of electrodes in the cerebellar cortex, currents of 22 microA mm-2 for 400 sec produced changes averaging -42% for inward current and +66% for outward current. The [K+]o changes near the surface were most rapid immediately after the onset of current and more gradual after some tens of seconds. Deeper within the tissue the rate of change was more uniform and after the end of stimulation the return to base line was slower. The amplitude, depth dependence and time course of the [K+]o changes were in reasonable agreement with the results calculated for a model in which K+ moves partly through extracellular space but primarily through membranes and cytoplasm within the tissue. The [K+]o changes were not attributable to variations in neuronal activity, although unit activity could be modified by current, since alternating currents failed to produce [K+]o changes and neither 0.1 mM-tetrodotoxin nor 5 mM-Mn2+ abolished the changes. The [K+]o changes were not abolished by topically applied ouabain (4 X 10(-4) M), 2,4-dinitrophenol (20 mM) or iodoacetate (10 mM), or by asphyxiation. Consequently the [K+]o changes are not dependent on metabolism. The data suggest that there is a selective mechanism for passive K+ transport in an electrochemical gradient within brain tissue that results in higher K+ fluxes than could be supported by ionic mobility in the extracellular fluid. This mechanism exists not only at the surface but within the brain parenchyma and may involve current flow through glial cells.