The effects of sea anemone toxin ATXII on single sodium channels were studied in cell-attached patches on rabbit ventricular myocytes at 20–22°C. Exposure of patches to 1,000 nM ATXII induced long-lasting bursts of openings, which were more dramatically different from control at −20 mV than at −50 mV. Mean open duration, which had a biphasic dependence on voltage in control patches, was monotonically dependent on voltage in toxin-exposed patches, being 3.5 times longer than control at −20 mV and 4.5 times longer at −10 mV. Multiple mean open durations were detected at depolarized potentials. To test whether the multiple mean open durations resulted from a mixture of modified and unmodified openings, histograms of late openings (when unmodified channels would be inactivated) were constructed. Because in most cases these fit a single exponential with a mean open duration like that of modified channels, we conclude that voltage-dependent toxin unbinding produced a mixed population of unmodified and modified openings. Consistent with this hypothesis, lower concentrations of toxin most often produced open-duration histograms best fit with two exponentials. Ensembles revealed complex decay kinetics, which could be interpreted within the context of the toxin-induced increase in mean open duration and burst duration and the summation of modified and unmodified events. Analysis of the numbers of early versus late events at −20 mV for patches exposed to 20 nM, 100 nM, and 1,000 nM ATXII predicted the ED50for ATXII block to be 285 nM at this potential. Using a five-state Markovian model, the action of ATXII could be explained as a reduction of the open-to-inactivated rate constant without effect on inactivation from closed states or other rate transitions.