We examined fish assemblage responses to urban intensify gradients in two contrasting metropolitan areas: Birmingham, Alabama (BIR) and Boston, Massachusetts (BOS). Urbanization was quantified by using an urban intensity index (UII) that included multiple stream buffers and basin land uses, human population density, and road density variables. We evaluated fish assemblage responses by using species richness metrics and detrended correspondence analyses (DCA). Fish species richness metrics included total fish species richness, and percentages of endemic species richness, alien species, and fluvial specialist species. Fish species richness decreased significantly with increasing urbanization in BIR (r = -0.82, P = 0.001) and BOS (r = -0.48, P = 0.008). Percentages of endemic species richness decreased significantly with increasing urbanization only in BIR (r = - 0.71, P = 0.001), whereas percentages of fluvial specialist species decreased significantly with increasing urbanization only in BOS (r = -0.56, P = 0.002). Our DCA results for BIR indicate that highly urbanized fish assemblages are composed primarily of largescale stoneroller Campostoma oligolepis, largemouth bass Micropterus salmoides, and creek chub Semotilus atromaculatus, whereas the highly urbanized fish assemblages in BOS are dominated by yellow perch Perca flavescens, bluegill Lefomis macrochirus, yellow bullhead Ameiurus natalis, largemouth bass, pumpkinseed L. gibbosus, brown bullhead A. nebulosus, and redfin pickerel Esox americanus. Differences in fish assemblage responses to urbanization between the two areas appear to be related to differences in nutrient enrichment, habitat alterations, and invasive species. Because species richness can increase or decrease with increasing urbanization, a general response model is not applicable. Instead, response models based on species' life histories, behavior, and autecologies offer greater potential for understanding fish assemblage responses to urbanization. ?? 2005 by the American Fisheries Society.