1. The apparent reversal potential (Erev) of the pace-maker current (iK2) is found to depend on the experimental protocol used for its measurement. Evidence is presented showing that depolarizing (hyperpolarizing) pulses given before a test hyperpolarization used to determine Erev, shift Erev to more negative (positive) values. These shifts are opposite to those expected if the only effect of pre-pulses were to change the concentration of potassium in extracellular clefts ([K]c) via accumulation and depletion processes. 2. This effect is shown to be due to the fact that Erev is dependent on s0, the degree of activation of iK2 at the start of the test hyperpolarization. 3. When a suitable protocol is used, depletion of cleft K can be demonstrated to take place during a large hyperpolarization. Changes in the level of [K]c induced by pre-pulses must therefore also affect the Erev determination. 4. A simplified three-compartment model has been used to investigate how K accumulation and depletion can affect the time course of iK2, with particular reference to the problem of Erev determination. Computed examples show that the model is able to reproduce the main features of the time course of iK2 recorded near its reversal potential and the changes induced by pre-pulses on Erev measuremnet. By contrast, simulation on a linear cable model rules out the possibility that such results are due to voltage non-uniformity. 5. The three-compartment model predicts that the measured value of Erev differs from EK2 for two reasons: (1) when the recorded current trace is flat iK2 is still outward and decaying, and (2) the K equilibrium potential shifts to more negative values while the test hyperpolarization is applied. 6. The finding that Erev is directly affected by changes in s at the beginning of the test pulse is discussed in relation to the action of agents (such as Ca2+, H+, salicylate, adrenaline and ouabain) which are found to shift both the s00 curve and Erev.