We conducted experimental feeding trials with larval and juvenile Bullfrogs (Lithobates catesbeianus) and Rough-skinned Newts (Taricha granulosa) to assess the accuracy of the scaled mass index (SMI). A control group was fed and a treatment group was starved within a randomized block design. After each of three trials, amphibian tissues were analyzed for lipid, protein, and water content. Mean pretreatment wet weight and SMI of individuals of each species and body form, representing two populations of equal body condition, were similar between control and treatment groups. Starved animals, representing a population in poor condition, had a 17–32% lower SMI than fed animals. Scaled fat and protein or lean mass were strongly correlated (r = 0.85–0.99) with SMI compared with percentages of fat, protein, or lean mass (r = 0.08–0.60). The SMI accurately reflected amphibian energy stores, but the depletion of energy stores differed by species and body form, with tadpoles retaining fat and the other species and body forms depleting fat stores. In addition, factors that we controlled in the laboratory (e.g., hydration, gut fill, reproductive state) may alter mass-length relationships in the field, so we advise collecting some specimens for body composition analyses to ensure the accuracy of the SMI when used in other applications.