The stream continuum concept suggests that the physical structure of the stream channel coupled with the hydrological cycle and energy inputs results in a consistent pattern of community structure and function along a stream. I evaluated this concept in a headwater fish community along two physical gradient gradients: upstream to downstream and riffle to pool. Habitat diversity and volume increased from upstream to downstream, and from riffle to pool. Temporal variation in habitat diversity was greater upstream. Upstream, and in riffles, habitat volume tended to vary more with time. Fluctuations in rainfall regime caused annual variation in habitat, especially in volume. Benthic insect density was highest from autumn (October—November) through spring (May—June). Following emergence of adults in late spring, invertebrate densities were low in summer in areas with riparian vegetation, but were not lower where riparian vegetation was absent and stable substrates were present. Along a gradient of substrates from silt—sand to gravel—rock, insect production increased, as indicated by adults and pupae in the drift. Peak resource availability for insectivore—piscivore fishes occurred in late summer and autumn, due to increased abundance of young—of—the—year fish. Pool and raceway—pool habitat guilds and insectivore and insectivore—piscivore trophic guilds contained the largest number of species. Increases in species richness were primarily associated with the addition of deeper habitats. Species richness of the pool insectivore—piscivore guild was especially variable over time. Biomass in shallow areas consisted predominantly of generalized insectivores. In deep, stable habitats, generalized insectivores were replaced as the predominant trophic group by insectivore—piscivores and large benthic insectivores. Immigration of fish occurred in spring and autumn, the periods of highest resource availability. Immigration between midriver and headwater regions primarily involved older age classes (III+), and was associated with changes in flow regime, habitat structure, and seasonal dynamics of the resource base of particular trophic groups. Flow regime and habitat volume appeared to be important factors limiting immigration in autumn, especially in pool species. Habitat diversity (depth, current, and substrate; DCS) was significantly correlated with fish species diversity (FSD). However, considerable variation occurred in the relationship between the two variables, including: (1) FSD decreased in winter in shallow, less diverse habitats due to emigration, and increased in spring due to recolonization; (2) FSD increased in spring and autumn when resource availability increased; (3) FSD was least predictable from DCS in autumn, when flow regimes were low and large numbers of fish recruits were present; and (4) DCS did not predict FSD as accurately in temporally variable upstream areas where large numbers of small fish dominated the community, especially in areas with human disturbance. Young age groups (0—II) were primarily found in shallow, temporally variable areas upstream and in riffles. Relative growth rates were highest during summer months. However, growth rates did not increase as much between spring and summer as would have been predicted from the seasonal increase in water temperature, suggesting that resource limitation may have occurred during summer periods. Centrarchids had substantially higher growth rates than cyprinids during early life stages. Net production for age 0—II fish was highest in upstream and riffle areas because of high densities of young, generalized insectivores. Net production of insectivore—piscivores was highest in downstream and pool habitats. Development of large, stable pools and raceways resulted in decreased fish production due to shifts in age structure toward fewer, large individuals with slower relative growth rates (age III+). Temporal variation in reproductive success and survival of younger age groups (0—I) was associated with random variation in high flow regimes and appeared to be a major factor determining spatial and temporal variation in production. These patterns of fish community structure and function support the qualitative aspect of the stream continuum concept: consistent shifts in community organization are associated with spatial or temporal changes in channel morphology and resource availability. However, for the stream continuum concept to be useful as a quantitative predictor, the ultimate mechanism(s) regulating fish community organization need to be more rigorously established. I hypothesize that in upstream or riffle areas, where younger age classes predominate, recolonization dynamics, the effect of gradual changes in physical conditions on competitive interactions, and temporal variation in reproductive success, are more important than competitive exclusion and predation as determinants of community organization.