1. The cell-attached configuration of the patch-clamp technique was used to assess resting membrane potential and firing threshold of CA1 pyramidal cells and interneurones of rat hippocampal slices. 2. Resting potential was inferred from the reversal potential of voltage-gated K+ currents with symmetrical intracellular and pipette K+ concentrations. Its mean value was -74 +/- 9 mV for silent interneurones (mean +/- s.d.; n = 17) and -84 +/- 7 mV for silent pyramidal cells (n = 8). Spontaneous action currents occurred in thirteen out of thirty-two interneurones and two out of ten pyramidal cells. In active cells, membrane potential values fluctuated by up to 20 mV, due in part to the large hyperpolarizations that followed an action current. 3. Membrane potential values determined from K+ current reversal were 13 +/- 6 mV more hyperpolarized than those measured in whole-cell recordings from the same neurones (n = 8), probably due to a Donnan equilibrium potential between pipette and cytoplasm. 4. Firing threshold of silent cells was determined by elevating external K+ until action currents were generated, while membrane potential was monitored from the cell-attached K+ current reversal. Spike threshold was attained at -49 +/- 8 mV for interneurones (n = 17) and at -60 +/- 8 mV for pyramidal cells (n = 8). Increasing external Ca2+ from 2 to 4 mM shifted the neuronal voltage threshold by +5 mV, without affecting resting potential. 5. For comparison with these values, we examined how the rate of membrane polarization influenced firing threshold in whole-cell records. Ramp current injections, of duration 15-1500 ms, revealed that current threshold followed a classical strength-duration relationship. In contrast voltage threshold, determined from current injection or by elevating extracellular K+, varied little with the rate of membrane polarization. 6. The state of activation and inactivation of Na+ and K+ currents might contribute to the stability of the voltage threshold. Cell-attached records showed that 79 +/- 10 % of Na+ channels and 64 +/- 10 % of K+ channels were available for activation at resting potential in silent cells (n = 8). As cells were depolarized to threshold, Na+ current availability was reduced to 23 +/- 10 %, and K+ current availability to 31 +/- 12 %. 7. The speed of transition into the inactivated states also appears to contribute to the invariance of threshold for all but the fastest depolarizations. At potentials close to threshold, the rate of inactivation of Na+ and K+ followed a double exponential time course, such that Na+ currents were 62 % inactivated and K+ currents were 63 % inactivated within 15 ms.