Several previous researchers reported that polymers- or gels can reduce permeability to water more than to oil. However, a plausible explanation for the phenomenon is not yet available. This property is critical to the success of gel treatments in production wells if zones cannot be isolated during gel placement. We examined how different types of gels reduce oil and water permeabilities in Berea sandstone. The gel formulations that we investigated included (1) resorcinol-formaldehyde, (2) Cr3+(chloride)-xanthan, (3) Cr3+(acetate)-polyacrylamide, and (4) colloidal silica. Several new methods were applied to obtain a better understanding of why gels can reduce water permeability more than oil permeability. First, before gel placement in cores, multiple imbibition and drainage cycles were performed in both flow directions. Results from these studies established that hysteresis of oil and water relative permeabilities was not responsible for the behavior observed during our subsequent gel studies. Second, several gels clearly reduced water permeability significantly more than oil permeability. Whereas previous literature reported this phenomenon for polymers and "weak" polymer-based gels, we also observed the disproportionate permeability reduction with a monomer-based gel (resorcinol-formaldehyde), as well as with both "weak" Cr3+(chloride)-xanthan and "strong" Cr3+(acetate)-HPAM gels. In contrast, a colloidal-silica gel reduced water and oil permeabilities by about the same factor. Residual resistance factors for several gels were found to erode during multiple cycles of oil and water injection. In spite of this erosion, the disproportionate permeability reduction persisted through the cycles for most of the gels. Studies using both oil and water tracers provided insight into the fraction of the pore volume occupied by gel. The strongest gels appeared to encapsulate the- original residual oil saturation—thus rendering the residual oil inaccessible during subsequent oil flooding.