The development of field sampling designs that employ multiple reference and polluted sites has been proposed as an alternative to the traditional upstream vs. downstream approach used in most biomonitoring studies. Spatially extensive monitoring programs can characterize ecological conditions within an ecoregion and provide the necessary background information to evaluate future changes in water quality. We measured physicochemical characteristics, heavy-metal concentrations, and benthic macroinvertebrate community structure at 95 sites in the Southern Rocky Mountain ecoregion in Colorado, USA. Most sites (82%) were selected using a systematic, randomized sampling design. Each site was placed into one of four metal categories (background, low, medium, and high metals), based on the cumulative criterion unit (CCU), which we defined as the ratio of the instream metal concentration to the U.S. Environmental Protection Agency criterion concentration, summed for all metals measured. A CCU of 1.0 represents a conservative estimate of the total metal concentration that, when exceeded, is likely to cause harm to aquatic organisms. Although the CCU was less than 2.0 at most (66.3%) of the sites, values exceeded 10.0 at 13 highly polluted stations. Differences among metal categories were highly significant for most measures of macroinvertebrate abundance and all measures of species richness. We observed the greatest effects on several species of heptageniid mayflies (Ephemeroptera: Heptageniidae), which were highly sensitive to heavy metals and were reduced by >75% at moderately polluted stations. The influence of taxonomic aggregation on responses to metals was also greatest for mayflies. In general, total abundance of mayflies and abundance of heptageniids were better indicators of metal pollution than abundance of dominant mayfly taxa. We used stepwise multiple-regression analyses to investigate the relationship between benthic community measures and physicochemical characteristics at the 78 randomly selected sites. Heavy-metal concentration was the most important predictor of benthic community structure at these sites. Because of the ubiquitous distribution of heavy-metal pollution in the Southern Rocky Mountain ecoregion, we conclude that potential effects of heavy metals should be considered when investigating large-scale spatial patterns of benthic macroinvertebrate communities in Colorado's mountain streams.