1The picture emerging from the rapidly growing literature on host-associated micro-2 biota is that host traits and fitness often depend on complex and interactive effects 3 of host genotype, microbial interactions, and abiotic environment. However, testing 4 these main and interactive effects typically requires large, multi-factorial experiments 5 and thus remains challenging in many systems. Furthermore, most studies of plant 6 microbiomes focus on terrestrial hosts and microbes. Aquatic habitats may confer 7 unique properties to plant micriobiomes. We grew different populations of duck-8 weed (Lemna minor ), a floating aquatic plant of increasing popularity in freshwater 9 phytoremediation, in three microbial treatments (adding no, "home", or "away" 10 microbes) at two levels of zinc, a common water contaminant in urban areas. Thus, 11 we simultaneously manipulated plant source population, microbial community, and 12 the abiotic environment, and measured both plant and microbial performance as 13 well as plant traits. Although we found little evidence of interactive effects, we found 14 strong main effects of plant source, microbial treatment, and zinc on both duckweed 15 and microbial growth, with significant variation among both duckweed and microbial 16 communities. Despite strong growth alignment between duckweed and microbes, zinc 17 consistently decreased plant growth, but increased microbial growth. Furthermore, 18 as in recent studies of terrestrial plants, microbial interactions altered a duckweed 19 phenotype (frond aggregation). Our results suggest that the duckweed source pop-20 ulation, its associated microbiome, and the contaminant environment may all need 21 to be considered in real-world phytoremediation efforts. Lastly, we propose that 22 duckweed microbes offer a robust experimental system for study of host-microbiota 23 interactions under a range of environmental stresses.24