Electrical Resistivity Tomography is a technique widely used for the investigation of the structure and fluid-dynamics of the shallow subsurface, particularly for hydro-geophysical purposes and using cross-borehole configurations . The results of ERT inversion and their usefulness in tackling and solving hydrogeophysical problems, even though invariably limited by resolution issues depends strongly on the accuracy of inversion, which in turns depends on a proper estimation and handling of data and model errors. Among model errors, one approximation often applied in cross-hole ERT and rarely evaluated in its consequences consists of neglecting the effects of boreholes and the fluids therein, that inevitably impact the current and potential patterns as measured by electrodes in the boreholes themselves. Few studies have pointed out such problems yet demonstrating that, particularly in presence of very saline fluids, this model approximation may prove inadequate and give rise to inversion artefacts. E.g. Osiensky et al. (2004) and Nimmer et al. (2008) demonstrated that apparent resistivity values in crosswell ERT experiments are significantly influenced by the borehole fluids. Unless the effects of the borehole fluids are accounted for, either by explicitly including the boreholes in the finite-element mesh or possibly by applying correction factors to the raw data, common regularized tomographic inversions are likely to yield images contaminated with artefacts. Some of the artefacts will be obvious - e.g., anomalous features along the lengths of the boreholes - whereas others might not be that easy to identify. In this contribution we present two case studies where the presence of very saline waters has a significant impact on the quality and significance of ERT inversions. In both cases the code R2 has been used (A. Binley, Lancaster University).