86Rb efflux from pinched-off rat brain presynaptic nerve terminals (synaptosomes) was used to measure the K permeability of the terminals. Synaptosomes were pre-loaded with 86Rb and the suspensions were then filtered on glass fibre filters. The terminals trapped on the filters were superfused with 'efflux solutions', and the effluent and filters were then counted. 86Rb efflux into physiological saline (PSS) containing 5 mM-K and 145 mM-Na was about 0.4% of the 86Rb load per second (component 'R'). Increasing extracellular K concentration [( K]o), or adding veratridine and sea anemone toxin, stimulated efflux; presumably by depolarizing the nerve terminals. The K-stimulated 86Rb efflux was a graded function of [K]o. High [K]o evoked at least three components of efflux: a 'fast phase' (T) that apparently inactivated in less than 1 s, a 'slower phase' (S) that was linear for 3-5 s, and a Ca-dependent phase (C). Some, but not all, of the slow phase 86Rb efflux (component S) may be attributable to increased efflux mediated by the 'resting' K permeability mechanism when the driving force is increased by depolarization. K efflux was also studied and was found to be qualitatively similar to 86Rb efflux. 86Rb: 42K permeability ratios were 0.6-0.8 for most components of the efflux. Raising the Mg concentration in the efflux solution shifted the 86Rb efflux versus [K]o curve in the direction of increased [K]o. This shift may be the result of screening of surface charges by Mg. Several agents that block various K channels in other preparations inhibited K-stimulated 86Rb efflux in synaptosomes: tetraethylammonium (TEA), tetrabutylammonium (TBA), and 4-aminopyridine (4-AP). The fast component (T) of high [K]o-stimulated 86Rb efflux was selectively blocked by low concentrations of 4-AP (apparent half-maximal inhibition, KI = 0.1-0.2 mM); it was also blocked by TEA (KI = 0.6 mM) and TBA (KI = 0.8-1.0 mM). Dose-response curves for inhibition of component T by all three agents were monophasic. the slow component (S) of the K-stimulated 86Rb efflux was much less sensitive to all three agents, than was component T; the broad dose-response curves were consistent with the view that two (or more) different K conductances may contribute to component S.(ABSTRACT TRUNCATED AT 400 WORDS)