Secondary production estimates from several Rocky Mountain streams were used to test hypotheses about the effects of chronic metal contamination on insect populations and ecosystem processes. Quantitative samples of chemistry, habitat, and benthic insects were collected monthly during the ice-free period (May–November) from five 2^nd- to 3^rd-order streams that varied primarily in Zn contamination. Secondary production was estimated for the 19 dominant taxa using increment-summation, size-frequency, and P/B methods. Uncertainty was estimated by bootstrapping estimates of mean abundance, biomass, and cohort production intervals. Secondary production of metal-sensitive Heptageniidae (Rhithrogena robusta, Cinygmula spp., and Epeorus longimanus) was lower in lightly to moderately contaminated streams than in reference streams. Experiments were done to determine whether herbivore growth was influenced by food quality in contaminated streams. Growth estimates from field and microcosm experiments revealed that low mayfly production in contaminated streams was caused mostly by reduced population abundances. Production of predatory stoneflies was also lower in contaminated streams than reference streams. Estimates of the trophic basis of production revealed that, although the relative contribution to community production from various food sources was similar among streams, total production attributable to algae and animal prey declined in contaminated streams. Much of the reduction in herbivory in contaminated streams was the result of lower production of heptageniids, especially R. robusta. Assemblage and taxon-specific estimates of secondary production were sensitive to variation in metal contamination and indicated that relatively low metal concentrations may have ecosystem-wide consequences for energy flow.