1. The effect of Centruroides sculpturatus scorpion venom on single frog myelinated nerve fibres was studied. Sodium currents through the nodal membrane were measured under voltage-clamp conditions before and after exposure to venom in Ringer solution 1-5 mug/ml. for 1-3 min. 2. Centruroides venom brings about repetitive firing and increased membrane potential noise. Spontaneous firing was also observed. Eventually the nodal membrane becomes inexcitable following venom treatment. 3. Under voltage clamp with a step depolarization of the membrane potential, activation and inactivation of sodium currents turns on, reaches a peak within about 25 msec, and then declines over several hundred milliseconds. As the amplitude and duration of the depolarizing pulse are increased, the size of the venom-induced current that follows also increases. 4. The venom-induced current turns on exponentially with a time constant near the value of the time constant for recovery from inactivation, tau-h, at the resting membrane potential. A depolarizing pulse inactivates this new current component, while a hyperpolarizing pulse leads to a larger venom-induced current immediately after the hyperpolarization. Its time course and membrane potential dependence indicate that the venom-induced current is modulated by the sodium inactivation process. 5. The membrane potential dependence of sodium activation in some channels is shifted by 40-50 mV in the hyperpolarizing direction. Depolarization increases the proportion of channels with shifted activation gating by first-older kinetics. Following a depolarizing pulse the activation parameter, m-3, remains elevated for hundreds of milliseconds, allowing channels to reopen as recovery from inactivation occurs. 6. A kinetic model with normal inactivation gating and shifted activation gating in some channels accounts for the observed voltage-clamp currents and for the repetitive firing evoked by Centruroides venom. In the model normal channels are converted to channels with shifted activation gating by a voltage dependent reaction. 7. The results suggest limits to possible coupling between sodium channel activation and inactivation. Transitions of the inactivation parameter, h, can occur normally in channels with a shifted membrane potential dependence for activation.