We describe 2 artificial stream experiments that exposed aquatic insect communities to zinc (Zn), copper (Cu), and cadmium (year 2014) and to Zn, Cu, and nickel (year 2015). The testing strategy was to concurrently expose insect communities to single metals and mixtures. Single-metal tests were repeated to evaluate the reproducibility of the methods and year-to-year variability. Metals were strongly accumulated in sediments, periphyton, and insect (caddisfly) tissues, with the highest concentrations occurring in periphyton. Sensitive mayflies declined in metal treatments, and effect concentrations could be predicted effectively from metal concentrations in either periphyton or water. Most responses were similar in the replicated tests, but median effect concentration values for the mayfly Rhithrogena sp. varied 20-fold between the tests, emphasizing the difficulty comparing sensitivities across studies and the value of repeated testing. Relative to the single-metal responses, the toxicity of the mixtures was either approximately additive or less than additive when calculated as the product of individual responses (response addition). However, even less-than-additive relative responses were sometimes greater than responses to similar concentrations tested singly. The ternary mixtures resulted in mayfly declines at concentrations that caused no declines in the concurrent single-metal tests. When updating species-sensitivity distributions (SSDs) with these results, the mayfly responses were among the most sensitive 10th percentile of available data for all 4 metals, refuting older literature placing mayflies in the insensitive portion of metal SSDs. Testing translocated aquatic insect communities in 30-d artificial streams is an efficient approach to generate multiple species effect values under quasi-natural conditions that are relevant to natural streams.