Conventionally 3 M KCl-filled, high resistance glass microelectrodes introduce large errors in the measurement of d.c. transmembrane potentials (MP) because of normally occurring tip potentials (TP). If, however, the 3 M KCl filling solution is adjusted to pH 2.0 with HCl, electrodes loose, almost entirely, the TP. Furthermore, the sensitivity of TP to the ionic composition of the test medium (0.125 M KCl and 0.125 M NaCl solution) is lost. Conventionally 3 M KCl-filled electrodes (pH 6.5) are found to have TP exceeding −30 mV in 0.125 M KCl solution. TP values measured in 0.125 M NaCl solution are about 29% higher than those measured in 0.125 M KCl solution. The MP of smooth muscle cells (isolated taenia coli of guinea-pig in modified Krebs solution at 32°C) determined using acidified electrodes shows an arithmetic mean of −51.2 mV with a S.D. of only 1.8 mV. However, MP determined by conventionally filled electrodes appears to be smaller by a variable amount. This reduction of MP is found to be linearly correlated over a wide range with the TP difference (Δ TP), as measured in the two test solutions. When comparing the MP determined by the two kinds of electrodes there is a good correlation provided that Δ TP is comparable. The results indicate that the reduction of MP determined by conventional electrodes is induced by their different TP in solutions of varying ionic composition. As a result, avoiding large TP by acidifying the filling solution can help to reduce TP-induced errors in MP measurements.