The effects that wastewater quality and mode of operation have on the performance of an asymmetric, hollow fiber, polysulfone, ultrafiltration (UF) membrane with a molecular weight cutoff of 100,000 Daltons were investigated. Performance was assessed through monitoring membrane flux, transmembrane pressure, effluent biochemical oxygen demand, and operational cost of the experimental system while treating filtered secondary, secondary, and filtered primary effluents. Fluxes achieved for filtered secondary (129-173 l/m2 h), secondary (101-158 l/m2 h), and filtered primary (20-41 l/m2 h) effluents were compared to those obtained at three other locations where similar UF systems were operated. A conceptual model of the impact of an insufficient backwash and of operating the UF system at constant flux on membrane performance is presented to explain the differences in fluxes. Employing pre-membrane granular filtration to remove a portion of the problematic particles in secondary effluent prior to UF led to optimal operational conditions. The costs associated with the operation of pre-membrane granular filtration were offset by the increase in production achieved. Although the use of recirculation could increase maintainable flux when treating a concentrated feed (e.g., filtered primary effluent), the associated costs were high. Improved UF performance was found to result from allowing flux to decline naturally, rather than using a constant flux mode of operation. The effluents produced when filtered secondary and secondary effluents were the feeds would be equivalent to an oxidized, coagulated, clarified, and filtered wastewater as per Title 22 California Wastewater Reclamation Criteria.