Fluid flow in unsaturated, fractured rock is studied with respect to applied environmental problems ranging from remediation of existing contaminated sites to evaluation of potential sites for isolation of hazardous or radioactive wastes. Spatial scales for such problems vary from meters to kilometers with temporal scales from months to tens of thousands of years. Because such scales often preclude direct physical exploration of system response and detailed site characterization, we are regularly forced to use our understanding (or misunderstanding) of the underlying physical processes to predict large scale behavior. It is essential that conceptual models used as the basis for prediction be firmly grounded in physical reality. In this paper, we provide examples of how recent advances in understanding of small-scale processes within discrete fractures may influence the behavior of fluid flow in fracture networks and ensembles of matrix blocks sufficiently to impact the formulation of intermediate-scale effective media properties. We also explore, by means of a thought experiment, how these same small-scale processes could couple to produce a large-scale system response inconsistent with current conceptual models of flow through unsaturated, fractured rock. ¿ American Geophysical Union 1995