A 3-year study of Wainiha River on Kaua'i, Hawai'i, was carried out to determine the impact that water removal had on key stream ecosystem parameters and functions. The study area included a diversion dam for a hydroelectric plant that removes water at an elevation of 213 m and returns it to the stream about 6 km downstream at an elevation of 30 m. There were two highelevation sites, one with undiverted flow and one with reduced flow, and two low-elevation sites, one with reduced flow and one with full flow restored. Monthly samples were taken of instream and riparian invertebrates and plants. When samples from similar elevations were compared, dewatered sites had lower concentrations of benthic photosynthetic pigments than full-flow sites, and benthic ash-free dry mass (AFDM) was higher at the two low-elevation sites regardless of flow. Benthic chlorophyll a (chl a) and AFDM were higher in summer months than in the winter. Benthic invertebrate abundance was highest at the full-flow, low-elevation site and benthic invertebrate biomass was highest at the full-flow, high-elevation site. Season had only marginal effects on abundance and biomass of benthic invertebrates. Diversity of benthic invertebrates was higher at the more-downstream sites. Abundance of drifting invertebrates was highest at the site above the diversion dam and generally higher in winter than in summer months. Biomass of drifting invertebrates was also highest at the above-dam site but there was little seasonal difference. Almost all parameters measured were lowest at the site just downstream of the diversion dam. The biotic parameters responded only weakly to flows that had occurred up to 1 month before the measurements were made. Flow, elevation, and season interact in complex ways that impact ecosystem parameters and functions, but water diversion can override all these environmental factors.In the history of limnology, the study of rivers (lotic systems) has lagged behind that of lakes (lentic systems). By the first decades of the twentieth century the basic understanding of the dynamics of thermal stratification provided a common framework for numerous subsequent studies of lentic systems (Ruttner 1953, Hutchinson 1957. For a long time students of rivers had no such unifying theme: hydrologists studied channel formation and, more practically, flood control, and biologists tended to focus on game fish or the production of food for these fish (see Hynes 1970 and Cummins et al. 1995 for historical accounts). Realization of the importance of complex interactions between the geomorphological setting of the stream and the physical and biological properties of the stream system set the stage for a conceptual model of stream dynamics based on func-