1. Intracellular recording techniques were used to characterize monosynaptic inhibitory postsynaptic potentials (IPSPs) and currents (IPSCs) in rat hippocampal slices and to study the mechanism of paired-pulse depression of these synaptic responses. This was achieved by stimulation in stratum radiatum close (less than 0.5 mm) to an intracellularly recorded CA1 neurone after pharmacological blockade of all excitatory synaptic transmission. 2. Under these conditions, low-frequency stimulation (0.033 Hz) evoked a pure biphasic IPSP, which had a short and constant latency to onset. This IPSP was blocked by tetrodotoxin (1 microM) suggesting that it resulted from the electrical stimulation of the axons and/or cell bodies of a monosynaptic inhibitory pathway. 3. Picrotoxin (100 microM) abolished the early component of the biphasic IPSP/C. It left an intact, pure late IPSP/C (IPSP/CB) which had a latency to onset of 29 +/- 2 ms, latency to peak of 139 +/- 4 ms, a duration of 723 +/- 135 (range 390-1730) ms and a reversal potential of -93 +/- 2 mV. The duration was highly dependent on the stimulus intensity whereas the latency to onset was largely independent of the stimulus intensity. The IPSP/CB was reduced or abolished by 1 mM-phaclofen. 4. Phaclofen (1 mM) and 2-hydroxy-saclofen (0.1-1.0 mM) reversibly depressed (60-100%) the late component of the biphasic IPSP/C and, where maximally effective, left a pure, early IPSP/C (IPSP/CA). The IPSP/CA had a latency to onset of 3 ms or less, a latency to peak of 17 +/- 1 ms, a duration of 225 +/- 3 ms and a reversal potential of -75 +/- 2 mV. 5. Two shocks of identical strength were applied in close succession to characterize, and to study the mechanisms underlying, frequency-dependent depression of inhibitory synaptic responses. Paired-pulse depression was seen for both phases of the biphasic IPSP/C and of the pure IPSP/CB, recorded in the presence of picrotoxin. Paired-pulse depression was not accompanied by changes in the reversal potential of either component, indicating that it was caused by a reduction in the two synaptic conductances. Paired-pulse depression was greater when high stimulus intensities were employed. 6. Paired stimuli were applied at separation intervals of between 5 ms and 10 s to determine the temporal profile of frequency-dependent depression. Paired-pulse depression of both IPSCA and IPSCB was most pronounced at an interstimulus interval of 100-125 ms and ceased to occur at intervals greater than 5 10s.(ABSTRACT TRUNCATED AT 400 WORDS)