1. Electrophysiological techniques are described which allow intracellular recording from peripheral myelinated axons of lizards and frogs for up to several hours. The sciatic and intramuscular axons studied here have resting potentials of -60 to -80 mV and action potentials (evoked by stimulation of the proximal nerve trunk) of 50-90 mV. They show a prominent depolarizing afterpotential (d.a.p.), which is present both in isolated axons and in axons still attached to their peripheral terminals. This d.a.p. has a peak amplitude of 5-20 mV at the resting potential, and decays with a half-time of 20-100 msec.2. The peak amplitude of the d.a.p. is voltage-sensitive, increasing to up to 26 mV with membrane hyperpolarization. The d.a.p. disappears as the axon is depolarized to -60 to -45 mV, and does not appear to reverse with further depolarization.3. The d.a.p. is not reduced when bath Ca is replaced by 2-10 mm divalent Mn or Ni. The d.a.p. is not reversed when axons depleted of Cl (by prolonged exposure to Cl-deficient, SO(4)-enriched solutions) are bathed in Cl-rich solutions. These results suggest that the d.a.p. is not mediated by a conductance change specific for Ca or Cl ions. Partial substitution of tetramethylammonium for bath Na, or addition of 10(-5)m-tetrodotoxin to the normal bathing solution, reduces the amplitude of both the action potential and the d.a.p.4. The amplitude of the d.a.p. is not sensitive to bath [K] over the range 1-7.5 mm, provided that all measurements are made at the same holding potential. This result argues that the d.a.p. is not mediated by accumulation of K outside the active axon.5. Treatments expected to inhibit the Na-K exchange pump (cooling from 25 to 10 degrees C, or 0.15 mm-ouabain) do not enlarge or prolong the d.a.p., although they do abolish a slower hyperpolarizing afterpotential seen following repetitive stimulation.6. The passive voltage response of the axon to small injected pulses of depolarizing or hyperpolarizing current shows a prominent, slowly decaying component with a time course similar to that of the d.a.p. Depolarizing current reduces the input resistance of the axon, and increases the rate of decay of both the passive voltage response and the d.a.p. There is a slight conductance increase during the peak of the d.a.p., but the same conductance increase can be produced by a comparable passive depolarization.7. We conclude that the d.a.p. is due mainly to a passive capacitative current, probably resulting from discharge of the internodal axonal membrane capacitance through a resistive current pathway beneath or through the myelin sheath. We suggest that this slow capacitative discharge becomes evident as soon as most of the nodal ionic channels activated during the action potential have closed. An electrical model of the myelinated axon that incorporates the postulated internodal leakage pathway can account both for the prolonged d.a.p. recorded inside the axon, and for the potential profile recorded extra-axonally in or near the internodal periaxonal space.