Summary The relative importance of active and passive transport processes in the clearing of potassium released from active neurons was estimated. Extracellular potassium activity [K +] 0 was measured with ion-selective microelectrodes in the sensory area of the neocortex and in lumbosacral spinal cord of cats. Transient elevation of [K +] 0 was evoked in cortex by stimulation of VPL and in spinal cord by stimulation of afferent nerves. The rate which excess [K +] 0 was cleared was either feebly or not at all influenced by variation of the intensity and frequency of stimulation. The half-decay times of [K +] 0 were however prolonged when the duration of stimulus trains was increased. Only small differences were seen in the rate of decay of [K +] 0 transients recorded at different locations within the gray matter; the shortest half-decay times occured where K + responses were largest. The different profiles of distribution of Δ [K +] 0 in response to stimulation of the cortical surface and of VPL nucleus were mapped. As in spinal cord also in cortex the distribution of the evoked sustained shifts of electric potential mirrored the distribution of [K +] 0 transients. The rate at which K + could diffuse out of volume sources similar in magnitude to the volumes of distribution of [K +] 0 responses in gray matter were calculated. The observed half-decay times of [K +] 0 transients were more than a hundred times shorter than those calculated for diffusion either in spinal cord or in cortex. Intravenous administration of digitoxigenin was shown to retard the clearing of [K +] 0 and caused an elevation of the unstimulated [K +] 0 baseline. Seizures were frequently induced by digitoxigenin when the [K +] 0 baseline was only slightly elevated, and the occurr3nce of seizures was not associated with a definable threshold level of [K +] 0. It is concluded that active reuptake is the principal mechanism of the clearing of [K + 0 released by neurons. Redistribution of K + by diffusion must have been negligible under the conditions of these experiments, but may be more important when only a few neurons release K + amongst many inactive cells. Considerations of a glial transport network are probably inconsequential for theories of the generation of seizures.