1. Intracellular recordings were obtained from anatomically verified layer II pyramidal cells in slices from rat piriform cortex cut perpendicular to the surface. 2. Responses to afferent and association fiber stimulation at resting membrane potential consisted of a depolarizing potential followed by a late hyperpolarizing potential (LHP). Membrane polarization by current injection revealed two components in the depolarizing potential: an initial excitatory postsynaptic potential (EPSP) followed at brief latency by an inhibitory postsynaptic potential (IPSP) that inverted with membrane depolarization and truncated the duration of the EPSP. 3. The early IPSP displayed the following characteristics suggesting mediation by gamma-aminobutyric acid (GABA) receptors linked to Cl- channels: associated conductance increase, sensitivity to increases in internal Cl- concentration, blockage by picrotoxin and bicuculline, and potentiation by pentobarbital sodium. The reversal potential was in the depolarizing direction with respect to resting membrane potential so that the inhibitory effect was exclusively via current shunting. 4. The LHP had an associated conductance increase and a reversal potential of -90 mV in normal bathing medium that shifted according to Nernst predictions for a K+ potential with changes in external K+ over the range 4.5-8 mM indicating mediation by the opening of K+ channels and ruling out an electrogenic pump origin. 5. Lack of effect of bath-applied 8-bromoadenosine 3',5'-cyclic monophosphate (8-Br-cAMP) or internally applied ethylene glycol-bis(beta-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA) on the LHP and failure of high amplitude, direct membrane depolarization to evoke a comparable potential, argue against endogenous mediation of the LHP by a Ca2+ activated K+ conductance [gK(Ca)]. However, an apparent endogenously mediated gK(Ca) with a duration much greater than the LHP was observed in a low percent of layer II pyramidal cells. Lack of effect of 8-Br-cAMP also indicates a lack of dependence of the LHP on cAMP. 6. Other characteristics of the LHP that were demonstrated include: a lack of blockage by GABAA receptor antagonists, a probable voltage sensitivity (decrease in amplitude in the depolarizing direction), and an apparent brief onset latency (less than 10 ms) when the early IPSP was blocked by picrotoxin. The LHP was unaffected by pentobarbital sodium when the early IPSP was blocked by picrotoxin. 7. Both the LHP and early IPSP were blocked by low Ca2+/high Mg2+, consistent with disynaptic mediation.(ABSTRACT TRUNCATED AT 400 WORDS)