Metasedimentary and metavolcanic rocks at Broken Hill, Australia, show regional-scale lowering of 8•sO values from as high as 16 per mil in Paragon Group metasedimentary rocks to values as low as 7 per mil within a few hundreds of meters of Pb-Zn-Ag orebodies. Such large-scale oxygen isotope resetting cannot be achieved by closed-system processes (such as partial melting or devolatization), implying that the rocks were affected by fluid flow. The preservation of peak metamorphic •sO fractionations between coexisting minerals, and a lack of correlation between 81sO values and the intensity of retrogression, suggests that oxygen isotope resetting occurred at, or prior to, the peak of regional metamorphism. Fluid flow during the peak of regional metamorphism is unlikely due to widespread fluid-absent partial melting and internal buffering of volatile activities at that time. Thus, fluid flow most probably predated regional metamorphism. The association of rocks with low 8•sO values with the Pb-Zn-Ag orebodies suggests a link with mineralization. The 8•sO values of rocks adjacent to the orebodies are similar to those recorded in volcanic-hosted massive sulfide deposits where convective circulation of ocean water has occurred. Base metal mineralization at Broken Hill may have occurred at hydrothermal vents at, or close to, the sea floor producing localized low 8•sO values with later kilometer-scale fluid circulation, possibly driven by the same or related intrusions, eausing regional resetting of oxygen isotopes. However, mineralization in skarn systems associated with pre-regional metamorphic granites is also a plausible model that could explain the regional 8•sO trends and much of the other geochemical data. Alternatively, the large-scale resetting of oxygen isotopes may be unrelated to mineralization. While the stable isotope data may not unambiguously constrain the origin of the base metal orebody, they indicate that mineralization most probably occurred prior to regional metamorphism.