Oocytes from Xenopus laevis are commonly used as an expression system for ion channel proteins. The most common method for their electrophysiological investigation is the two-microelectrode voltage clamp technique. The quality of voltage clamp recordings obtained with this technique is poor when membrane currents are large and when rapid charging of the membrane is desired. Detailed mathematical modeling of the experimental setup shows that the reasons for this weak performance are the electrical properties of the oocytes and the geometry of the setup. We measured the cytosolic conductivity to be approximately 5 times lower than that of the typical bath solution, and the specific membrane capacitance to be approximately 6 times higher than that of a simple lipid bilayer. The diameter of oocytes is typically approximately 1 mm, whereas the penetration depth of the microelectrodes is limited to approximately 100 microm. This eccentric current injection, in combination with the large time constants caused by the low conductivity and the high capacitance, yields large deviations from isopotentiality that decay slowly with time constants of up to 150 micros. The inhomogeneity of the membrane potential can be greatly reduced by introducing an additional, extracellular current-passing electrode. The geometrical and electrical parameters of the setup are optimized and initial experiments show that this method should allow for faster and more uniform control of membrane potential.