1. The nature, distribution and function of rectifying channels in rat spinal root myelinated axons has been assessed with selective blocking agents and a variety of intracellular and extracellular recording techniques. 2. The electrotonic responses of roots poisoned with tetrodotoxin (TTX) to constant current pulses had fast (rise time much less than 1 ms) and slow components, which were interpreted in terms of Barrett & Barrett's (1982) revised cable model for myelinated nerve. Depolarization evoked a rapid outward rectification (time constant, tau approximately 0.5 ms), selectively blocked by 4-aminopyridine (4AP, 1 mM), and a slow outward rectification (tau approximately 15 ms), selectively blocked by tetraethylammonium (TEA, 1 mM) or Ba2+ (0.5 mM). Hyperpolarization evoked an even slower inward rectification, selectively blocked by Cs+ (3 mM) but not by Ba2+. 3. From the different effects of the blocking agents on the fast and slow components of electrotonus, it was deduced (a) that the inward rectification is a property of the internodal axon, (b) that the slow outward rectifier is present at the nodes, and probably the internodes as well, and (c) that the 4AP-sensitive channels have a minor nodal and a major internodal representation. 4. TEA and Ba2+ reduced the accommodation of roots and fibres not poisoned with TTX to long current pulses, whereas 4AP facilitated short bursts of impulses in response to a single brief stimulus. 5. TEA and Ba2+ also abolished a late hyperpolarizing after-potential (peaking at 20-80 ms), while 4AP enhanced the depolarizing after-potential in normal fibres, and abolished an early hyperpolarizing after-potential (peaking at 1-3 ms) in depolarized fibres. Corresponding to the later after-potentials were post-spike changes in excitability and conduction velocity, which were affected similarly by the blocking agents. Cs+ increased the post-tetanic depression attributable to electrogenic hyperpolarization. 6. The physiological roles of the three different rectifying conductances are discussed. It is also argued that the prominent ohmic 'leak conductance', usually ascribed to the nodal axon, must arise in an extracellular pathway in series with the rectifying internodal axon.