1. The mode of action of various antagonists of acetylcholine (ACh) excitatory effects on Aplysia neurones was studied under voltage clamp. ACh was applied by iontophoresis whereas antagonists were applied in the bath. Tubocurarine and hexamethonium were the most thoroughly studied compounds. 2. The 'elementary current', calculated as the ratio of the variance of the ACh noise to the mean ACh induced current, was not modified by any of the antagonists tested. 3. The evolution of the ACh induced current after a voltage jump, which is normally described by a single exponential, was modified by all the antagonists tested. A common feature of the modified relaxations was the appearance, over a certain concentration range of the antagonist, of two successive and opposite exponential components. 4. The characteristics of the composite relaxations depend on the antagonist. For a given antagonist they vary with membrane potential, ACh concentration, and antagonist concentration. 5. The noise power spectra of the ACh induced current showed changes consistent with those of the relaxations. 6. In the absence of antagonists, the current induced by a steady application of ACh increases linearly with hyperpolarization. In the presence of antagonists, the I-V curve shows a marked curvature, indicating a proportionally larger reduction of the ACh response at more negative membrane potentials. 7. The voltage sensitivity of the blocking action of hexamethonium and decamethonium is noticeably stronger than that of monovalent antagonists. 8. A model is proposed which accounts for the observed effects. It assumes that the antagonists studied bind perferentially to the 'activated' ACh-receptor complex, and convert it to a non-conducting state. Kinetic constants can be calculated for this reaction; e.g. for curare, at 12 degrees C and -80 mV, the dissociation and association constants were estimated at 0.1 sec-1 and 4 X 10(5) M-1 sec-1. 9. Partial replacement of the extracellular Na by Tris modifies the relaxations observed in the presence of hexamethonium. Hexamethonium appears less effective in the presence of Tris, which supports the hypothesis that the binding site of the antagonists is linked to the ionic channel.