1. The putative neurotransmitter FMRFa (Phe-Met-Arg-Phe-amide) caused an inhibitory modulation of the voltage-gated sodium current (INa) in central neurones, the peptidergic caudo dorsal cells (CDCs) of the mollusc Lymnaea stagnalis. FMRFa reduced INa at all command potentials tested (ranging from -35 to +20 mV), but the amplitude of the effect of FMRFa was voltage dependent, inhibition being stronger at more negative potentials (50 +/- 5% reduction at half-maximal INa activation versus 25 +/- 8% at the peak of the I-V curve). 2. INa current traces were well fitted by a Hodgkin & Huxley based model, using m3 activation kinetics and two time constants for inactivation. 3. The steady-state inactivation curve of INa was characterized by half-maximal inactivation at -42.5 +/- 1.81 mV and a slope factor of 4.6 +/- 0.28 mV. The fastest time constant of inactivation ran from 100 +/- 5 to 0.8 +/- 0.32 ms and the slower time constant from 505 +/- 45 to 4.8 +/- 1.40 ms in the range -40 to -5 mV. 4. FMRFa had no significant effect on either component of inactivation, nor on the voltage dependence of steady-state inactivation, nor on the maximal conductance. 5. FMRFa affected the activation of INa. The activation time constant was increased, ranging from 0.75 +/- 0.050 to 0.22 +/- 0.017 ms under control and from 0.91 +/- 0.043 to 0.31 +/- 0.038 ms with FMRFa in the voltage range -25 to +5 mV. The steady-state activation curve was shifted to less negative potentials: half-maximal activation occurred at -26.5 +/- 1.2 mV under control and at 23.6 +/- 1.4 mV with FMRFa; the slope factor (4.6 +/- 1.4 mV in control experiments) was not affected. The combination of slower activation kinetics and a shift in the voltage dependence of activation in the Hodgkin & Huxley based model, adequately explained the reduction of INa by FMRFa. 6. The physiological consequence is that the spiking threshold is increased, causing an arrest of on-going firing activity and a decrease in excitability.