Premise Outcomes of species interactions, especially mutualisms, are notoriously dependent on environmental context, and environments are changing rapidly. Studies have investigated how mutualisms respond to or ameliorate anthropogenic environmental changes, but most have focused on nutrient pollution or climate change and tested stressors one at a time. Relatively little is known about how mutualisms may be altered by or buffer the effects of multiple chemical contaminants, which differ fundamentally from nutrient or climate stressors and are especially widespread in aquatic habitats. Methods We investigated the impacts of two contaminants on interactions between the duckweed Lemna minor and its microbiome. Sodium chloride (salt) and benzotriazole (a corrosion inhibitor) often co‐occur in runoff to water bodies where duckweeds reside. We tested three L. minor genotypes with and without the culturable portion of their microbiome across field‐realistic gradients of salt (3 levels) and benzotriazole (4 levels) in a fully factorial experiment (24 treatments, tested on each genotype) and measured plant and microbial growth. Results Stressors had conditional effects. Salt decreased both plant and microbial growth and decreased plant survival more as benzotriazole concentrations increased. In contrast, benzotriazole did not affect microbial abundance and even benefited plants when salt and microbes were absent, perhaps due to biotransformation into growth‐promoting compounds. Microbes did not ameliorate duckweed stressors; microbial inoculation increased plant growth, but not at high salt concentrations. Conclusions Our results suggest that multiple stressors matter when predicting responses of mutualisms to global change and that beneficial microbes may not always buffer hosts against stress.

1Premise Environments are changing rapidly, and outcomes of species interactions, es-2 pecially mutualisms, are notoriously dependent on the environment. A growing number 3 of studies have investigated responses of mutualisms to anthropogenic changes, yet most 4 studies have focused on nutrient pollution or climate change, and tested single stres-5 sors. Relatively little is known about impacts of simultaneous chemical contaminants, 6 which may di↵er fundamentally from nutrient or climate stressors, and are especially 7 widespread in aquatic habitats. 8 Methods We investigated the impacts of two common contaminants on interactions 9 between the common duckweed Lemna minor and its microbiome. Sodium chloride 10 (salt) and benzotriazole (a corrosion inhibitor) negatively a↵ect aquatic organisms 11 individually, yet commonly co-occur in runo↵ to duckweed-inhabited sites. We tested 12 three L. minor genotypes with and without the culturable portion of their microbiome 13 across field realistic gradients of salt (3 levels) and benzotriazole (4 levels) in a fully 14 factorial experiment (72 treatments), and measured plant and microbial growth. 15Key Results We found that stressors had conditional e↵ects. Salt decreased both 16 plant and microbial growth, but decreased plant survival more as benzotriazole con-17 centrations increased. In contrast, benzotriazole did not a↵ect microbial abundance, 18 and benefited plants when salt and microbes were absent, perhaps due to the biotrans-19 formation we observed without salt. Microbes did not ameliorate duckweed stressors, 20 as microbial inoculation increased plant growth, but not at high salt concentrations. 21Conclusions Our results suggest that multistressor e↵ects matter when predicting re-22 sponses of mutualisms to global change, but that mutualisms may not bu↵er organisms 23 from stressors. 24 dation 26 1 29the fates of interacting species, and most organisms engage in at least one mutualism with 30 another species, be it for nutrition, protection, or dispersal. There is concern that mutual-31 ists may be particularly vulnerable to environmental change because mutualism outcomes 32 are often context-dependent. In other words, the magnitude of benefits exchanged between 33 mutualistic partners typically depends on the external abiotic or biotic environment (Cham-34 berlain et al., 2014), and environments are changing at a historically unprecedented pace. As 35 a result, many mutualisms are now taking place in novel environments where partners face 36 multiple stressors (Kiers et al., 2015). A key question is therefore: how do anthropogenic 37 stressors, singly or in combination, a↵ect mutualism outcomes? And in the same vein, does 38 mutualism mitigate or exacerbate global change (Frederickson, 2017)? 39 There is also currently substantial interest in plant-associated microbiomes and how 40 they may facilitate acclimation or adaptation to plant stressors. The term 'microbiome' 41 refers either to the community of microbes living together in an environment, suc...