Responses of invertebrate assemblages along gradients of urban intensity were examined in three metropolitan areas with contrasting climates and topography (Boston, Massachusetts; Birmingham, Alabama; Salt Lake City, Utah). Urban gradients were defined using an urban intensity index (UII) derived from basin-scale population, infrastructure, land-use, land-cover, and socioeconomic characteristics. Responses based on assemblage metrics, indices of biotic integrity (B-IBI), and ordinations were readily detected in all three urban areas and many responses could be accurately predicted simply using regional UIIs. Responses to UII were linear and did not indicate any initial resistance to urbanization. Richness metrics were better indicators of urbanization than were density metrics. Metrics that were good indicators were specific to each study except for a richness-based tolerance metric (TOLr) and one B-IBI. Tolerances to urbanization were derived for 205 taxa. These tolerances differed among studies and with published tolerance values, but provided similar characterizations of site conditions. Basin-scale land-use changes were the most important variables for explaining invertebrate responses to urbanization. Some chemical and instream physical habitat variables were important in individual studies, but not among studies. Optimizing the study design to detect basin-scale effects may have reduced the ability to detect local-scale effects. ?? 2005 by the American Fisheries Society.