Almost all potassium channels within mammalian myelinated nerve fibers are covered by the myelin sheath and their majority is concentrated in a small paranodal region. In order to investigate effects of this paranodal potassium permeability on nerve fiber behavior via a simulation approach, a myelinated fiber model is required that treats myelin sheath and internodal axolemma as separate entities. Such a fiber description was developed by Blight (1985) and his model was used to investigate the effects paranodal potassium channels have on the ability of maintaining repetitive firing in response to a constant current injected into the fiber. It was found that increasing the potassium channel density at the paranode from low to moderate values widened the range of injected currents with a repetitive response. This promotion of repetitive activity by the introduction of additional potassium channels occurred up to an "optimal" value beyond which a further increase in paranodal potassium permeability narrowed the range of currents with a repetitive response. Finally, if a certain limit in paranodal potassium channel density was exceeded, repetitive activity was abolished completely. These results were obtained regardless of the assumptions about the electrical resistance of the myelin sheath. On the other hand, in the absence of potassium channels repetitive firing could be observed only when a high resistance myelin sheath was assumed, whereas a nerve fiber model with electrical properties inferred from intracellular recordings needed at least some potassium channels within the paranodal region for repetitive firing in response to an injected current.