Magnesium (Mg) is a mining-related contaminant in the Alligators Rivers Region of tropical northern Australia. A mesocosm experiment was used to assess Mg toxicity to aquatic freshwater assemblages. Twenty-five 2700-L tubs were arranged, stratified randomly, on the bed of Magela Creek, a seasonally flowing, sandy stream channel in the Alligator Rivers Region of northern Australia. The experiment comprised 5 replicates of 4 nominal Mg treatments, 2.5, 7.5, 23, and 68 mg L−1, and a control. Phytoplankton biomass, and diatom, zooplankton, and macroinvertebrate assemblages present in the treatment tubs were sampled before and after Mg addition. A significant negative relationship between phytoplankton biomass and Mg was observed 4 wk after Mg addition as measured by chlorophyll a concentrations (r2 = 0.97, p = 0.01). This result was supported by reductions in some major phytoplankton groups in response to increasing Mg concentrations, in the same experiment and from independent field studies. There was a significant negative relationship between zooplankton assemblage similarity (to control) and Mg concentrations (r2 = 0.96, p = 0.002). Seven weeks after Mg addition, macroinvertebrate assemblages were dominated by 3 microcrustacean groups (Ostracoda, Cladocera, and Copepoda), each reaching maximum abundance at intermediate Mg concentrations (i.e., unimodal responses). The responses of phytoplankton and zooplankton were used to derive assemblage effect concentrations (Mg concentrations resulting in x% of the assemblage change [ECx]). Magnesium concentrations resulting in assemblage EC01 values were <3 mg L−1. Together with candidate guideline values from other laboratory- and field-based lines of evidence, the mesocosm EC01 values were incorporated into a weight-of-evidence framework for a robust regulatory approach to environmental protection. Environ Toxicol Chem 2020;00:1–15.