Petrophysical information, including stress release data, recently made available for low-porosity (0.1--0.4%) crystalline rocks has been used to develop a ''broken conductor model'' that may explain the origin of enhanced electrical conductivity of the intermediate to lower continental crusts of shields and its disappearance when the rocks are uplifted. Average pore size and pore density of their fluid flow paths are 0.1--0.4 μm and 1--400/m, respectively. The model assumes that these paths are filled or lined with interconnected electrical conducting minerals, such as graphite, resulting in the rocks having resistivities as low as 40 to 800 &OHgr; m in situ. Graphite film thicknesses required to produce these resistivities are 5--20 nm (50--200 ¿). The model also assumes that electrical conducting paths break up during uplift due to stress release, resulting in a poorly conducting rock. Petrophysical data of granites and observations from deep drill holes suggest a porosity increase of the order of 3% from an uplift of 10 km or more. However, the newly formed pore space would fill with deposits or intruded material so that surface rock porosities would not be much larger than at depth. Electrical resistivities of rock samples containing disconnected grain boundary graphite presently exposed by the Kapuskasing uplift of the Canadian Shield are in the range of 3000--12000 &OHgr; m for the direction of foliation. These values are 10--1000 times larger than the estimates for similar rocks containing interconnected graphite but are 10--100 times smaller than usually observed in the Canadian Shield for similar rocks absent of grain boundary graphite. These observations are in good agreement with the results predicted by the Broken conductor model for an uplift of 10 km or more if a gross linear relationship is assumed between crack porosity and uplift for the upper 10 km of the crust. ¿ American Geophysical Union 1993