This extended abstract discusses the limitations of EPANET. Like most other distribution system water quality models, EPANET is limited to tracking the dynamics of a single component as it is transported throughout the network of pipes and storage tanks. Such models cannot consider interactions between bulk components that are transported with the water velocity, and surface components that are fixed to the pipe wall. These are serious limitations, since all water quality indices result from reactions between chemical or biological components that may be suspended or dissolved in bulk water, or attached or adsorbed to the pipe wall. A single specie model implies that distribution system water quality can not be described mathematically in a way that mimics the underlying mechanisms. It is a restrictive framework that precludes experimentation with alternative process models, and has limited research and development of network water quality models for the last 15 years. Aside from significant software and algorithmic improvements, today's distribution system water quality models have progressed little beyond those first reported in 1988 (Grayman, 1988). This extended abstract presents several examples that illustrate the need to move beyond single-specie water quality models, and suggests two EPANET water quality model extensions that could improve the technology. Includes extended abstract only.