1. Inward currents in chromaffin cells were studied with the patch-clamp technique (Hamill, Marty, Neher, Sakmann & Sigworth, 1981). The intracellular solution contained 120 mM-Cs(+) and 20 mM-tetraethylammonium (TEA(+)). Na(+) currents were studied after blockade of Ca(2+) channels with 1 mM-Co(2+) applied externally. Ca(2+) currents were recorded after eliminating Na(+) currents with tetrodotoxin (TTX). The current recordings were obtained in cell-attached, outside-out and whole-cell recording configurations (Hamill et al. 1981).2. Single channel measurements gave an elementary current amplitude of 1 pA at -10 mV for Na(+) channels. This amplitude increased with hyperpolarization between -10 and -40 mV, but did not vary significantly between -40 and -70 mV.3. The mean Na(+) channel open time was 1 ms at -30 mV. This open time decreased both with depolarization and hyperpolarization. Its value was close to the time constant of inactivation, tau(h), above -20 mV.4. Ensemble fluctuation analysis of Na(+) currents gave results consistent with those of single channel measurements. Noise power spectra obtained between -35 mV and 0 mV could be fitted with a single Lorentzian. A range of Na(+) channel densities of 1.5-10 channels per mum(2) was calculated.5. Cell-attached single Ca(2+) channel recordings were obtained in isotonic BaCl(2) solution. The single channel amplitude was 0.9 pA at -5 mV, and it became smaller for positive potential values.6. At -5 mV, single Ba(2+) currents appeared as bursts of 1.9 ms mean duration containing on the average 0.6 short gaps. The burst duration was larger at positive potentials.7. Ensemble fluctuation analysis of Ca(2+) channels was performed on whole-cell recordings in external solutions containing isotonic BaCl(2) or external Ca(2+) (Ca(o)) concentrations of 1 and 5 mM. The unit amplitude calculated in the former case was similar to that obtained in single channel measurements.8. Noise power spectra of Ca(2+) or Ba(2+) currents could be fitted by the sum of two, but not one, Lorentzian components.9. Tail currents could be fitted by the sum of two exponential components. The corresponding time constants had values close to those obtained with noise analysis.10. The rising phase of Ca(2+) and Ba(2+) currents was sigmoid. It could be fitted by the sum of three exponentials. The time constant of the largest amplitude component, tau(1), was similar to the time constants of the slow component observed in noise and tail experiments. This time constant also corresponded to the burst duration obtained in single channel measurements.11. The value of tau(1) was larger in 5 mM-Ca(o) and in isotonic Ba(2+) than in 5 mM-Ba(o). Thus, the kinetic properties of Ca(2+) channels depend on the nature and concentration of the permeating ion.12. A simple kinetic scheme is proposed to model the activation pathway of Ca(2+) channels.13. Currents in 1 mM-Ca(o) and 5 mM-Ca(o) showed clear reversals around +53 mV and +64 mV respectively. The outward currents observed above these potentials are most probably due to Cs(+) ions flowing through Ca(2+) channels.14. The instantaneous current-voltage relation was obtained from tail current data in the range -70 to +100 mV in 5 mM-Ca(o). The resulting curve displayed an inflexion point around the reversal potential.15. Very little inactivation of Ca(2+) currents was observed. However, a slow current decline was observed in some cells above +10 mV.16. Conditioning prepulses to positive potentials had potentiating or depressing effects on Ca(2+) currents depending on whether the test pulse lay below or above the maximal current potential. The potentiating effect may be linked to the slowest component of the current rise observed below +10 mV. The depressing effect may be related to the slow decline obtained above +10 mV.17. Analysis of ensemble variance and of tail current amplitudes suggested that the opening probability of Ca(2+) channels was at least 0.9 above +40 mV.18. A slow rundown of Ca(2+) currents was observed in whole-cell recordings. The speed of the rundown was dependent on intracellular Ca(2+) concentration. The rundown was apparently due to a progressive elimination of the channels available for activation.19. The density of Ca(2+) channels (before rundown) was estimated at 5-15/mum(2).20. In cell-attached experiments, inward current channels were often seen to follow action potentials. These events did not appear to be the usual Na(+) and Ca(2+) currents. They were probably due to cation influx of either Na(+) or Ba(2+), depending on the pipette solution, through Ca(2+)-dependent channels. Voltage-independent single channel activity observed in whole-cell and outside-out recordings may be due to the same channels.