The study of geothermal reservoir behavior is presently in a state of change brought about by the discovery that reservoir heterogeneity--fractures in particular--is responsible for large scale effects during production. On the other hand, some parts of a reservoir, or some portions of its behavior. may be unaffected by fractures and behave, instead, as if the reservoir were a homogeneous porous medium. Drilling has for many years been guided by geologists prospecting for fractures (which have been recognized as the source of production), but until recently reservoir engineers have not studied the behavior of fractured systems under production. In the last three years research efforts, funded by the Department of Energy and others, have made significant progress in the study of fractures. The investigations into simulation of fracture flow, tracer analysis of fractured systems, and well test analysis of double porosity reservoirs are all advancing. However, presently we are at something of a conceptual impasse in defining a reservoir as fractured or porous.