Lower input resistance with intracellular recording, rather than with whole cell recording, usually has been ascribed to a shunt produced by penetration injury. An alternative explanation is a higher input resistance during whole cell recording due to wash-out of cytoplasmatic substances. We have used neuronal polarization at the onset and termination of an applied electric field for shunt detection. An analytical expression was derived for field-induced polarization in a shunted ohmic cable. When the shunt is negligible, the transient response to a step in DC field decays much faster than the response to current injected through the recording electrode. In the case of a significant shunt an over- and undershoot of the transmembrane potential appear at the shunted end when the field is switched on and off. Over- and undershoot decay with the same slowest time constant as the response to injected current. The results for the cable are generalized for nonuniform fields and arbitrary branching neurons with homogeneous membrane. The field effect was calculated for two reconstructed neurons with different branching pattern. The calculations confirmed the theoretical inferences. The field polarization can be used for shunt detection. The theory was checked experimentally in 18 ventral neurons in transverse slices of the turtle spinal cord. In seven neurons, field-induced under- and overshoots were observed when sharp electrodes were used. This indicates the presence of an injury shunt. In the remaining 11 neurons, however, there were no under- or overshoots, indicating that a shunt is not always induced. When patch electrodes were used, the seal quality was checked by inducing a spike with a strong field stimulus before and after the rupture of the membrane. When the threshold field strength for spike initiation was not changed by membrane rupture, under- and overshoots were not observed. This was taken to indicate a good seal. In such recordings under- and overshoots were observed when a shunt was induced by local application of glycine. The fast and monotonic response to weak field stimulation suggests homogeneous electric properties of the soma-dendritic membrane when active conductances are not recruited. We propose using polarization by weak DC fields to ensure the quality of recordings with sharp and whole cell electrodes and for checking the ohmic homogeneity of the membrane. These controls are particularly important for evaluation of electrotonic parameters.