The three-dimensional induced-polarization and resistivity-modeling problem for buried source and receiver electrodes is solved by using a modified form of Barnett's surface-integral technique originally developed for surface-electrode configurations. Six different buried electrode configurations are considered in this study: three types of hole-to-hole configurations, hole-to-surface and surface-to-hole configurations, and the single hole (bipole-bipole) configuration. Results show there is no 'best' method for all situations encountered in the field. The choice of method depends upon depth of the body, spacing of drill holes, and electrical properties of the body.In hole-to-hole measurements, the geometric factor (necessary for the computation of the apparent resistivity) becomes infinitely large or infinitely small whenever the receiving bipole is placed at a depth so that it lies on a zero equipotential surface. This leads to the formation of apparent resistivity anomalies that are extremely sensitive to the presence of the body but that are also complicated and not easily correlated with the position of the body. It is shown that diagnostic and easily interpretable anomalies are obtained by selecting the proper source-receiver configurations.