1. Using single node preparations of the bull frog or the toad, observations were made on the variation of the voltage across the nodal membrane under various experimental conditions.
2. The time constant of the variation in the membrane voltage caused by a long subthreshold rectangular pulse was of the order of 0.1 msec.
3. The action potential was initiated when the potential inside the node was raised stimulating pulses above a threshold level of approximately 15 mv. for a node in normal Ringer; it was greater in a relatively refractory node and in a partially narcotized node.
4. The variation of the membrane voltage caused by long stimulating pulses of subrheobasic strengths was in general proportional to the strength of the applied pulse. A non-linear behavior of the membrane voltage was observed with barely subthreshold stimulating pulses.
5. The early portion of the action potential of a node was not modified by a direct current which was strong enough to produce measurable potential changes (IR drops) across the resting membrane.
6. A strong pulse of inward current applied to the node during activity abolished the portion of the action potential following the pulse in all-or-none manner.
7. There was no refractory period after a response abolished in its early phase. Following a response abolished later, the recovery in the spike height started from the level of the action potential at the time of abolition.
8. Initiation and abolition of action potentials at a single node are interpreted as "transitions" between the two "equilibrium potential levels" at the node.