An anodal pulse applied during the falling phase of an action potential, if weak, produces a slight enhancement of negativity of the falling phase, but if increased in amplitude produces a split of the action potential into an early and "delayed" response and finally, complete abolition of the falling phase. If the pulse amplitude is increased still more after abolition, a second response is elicited. The latency to this second response following abolition is shorter than the latency to the delayed response, and further increase of the applied pulse amplitude cannot abolish this second response. To obtain abolition of the delayed response it is necessary to apply a considerably stronger anodal pulse near the peak of the spike than later during the falling phase. The reverse is true to obtain the second response. The anodal pulse sufficient to produce anode break excitation during the action potential and elicit the second response is ineffective applied to a resting node membrane. It is postulated that: a) due to an effect of the action potential itself the membrane is being actively depolarized during the early falling phase of the spike and b) the excitability of the node membrane is actually retained both during and following an action potential in the so-called refractory period but requires "resetting" by a positive pulse in order for re-excitation to take place.