Metabolic rate is commonly thought to scale with body mass to the 3/4-power as a result of universal body-design constraints. However, recent comparative work has shown that the metabolic scaling slope may vary significantly among species and higher taxa, apparently in response to different lifestyles and ecological conditions, though the precise mechanisms involved are not well understood. To better understand these under-appreciated ecological effects and their causes, it is important to control for extraneous phylogenetic and environmental influences. We demonstrate how this may be done by comparing the ontogenetic scaling of resting metabolic rate among populations of the same species (the amphipod Gammarus minus) in mid-Appalachian freshwater springs with similar, relatively constant environmental conditions, except for the varying presence of the predatory fish Cottus cognatus. We found that populations of G. minus exhibit significantly lower metabolic scaling slopes (0.54 to 0.62) in three freshwater springs with C. cognatus than in two springs without these fish (0.76 to 0.77). We tested multiple hypothetical causes for these population differences. Our results best supported the hypothesis that metabolic scaling was influenced by the effects of size-selective predation on the ontogeny of growth, a metabolically expensive process. The body-size scaling of growth is significantly less steep in the populations inhabiting springs with versus without fish, thus paralleling the interpopulation differences in metabolic scaling. Prematurational growth of G. minus is as high or higher in the fish springs, whereas postmaturational growth is significantly lower, often approaching zero. Similarly the amphipods in the fish springs tend to have higher metabolic rates at small sizes, but lower metabolic rates at large sizes, compared to those in the fishless springs. Our results do not support other hypothetical causes of the interpopulation variation in metabolic scaling, including differential scaling of cell size or low-metabolism body components (fat and mineralized exoskeleton), or possible effects of other environmental factors associated with the presence of fish. However, fish-induced population differences in adult behavioral activity may influence metabolic scaling in G. minus, a possibility under current study. We conclude that ecological factors may significantly influence metabolic scaling, contrary to common belief.