Background/Question/Methods  Plant associated microbes both above- and below-ground have been increasingly recognized for their individual influence on host population dynamics. However, the degree to which above- and below-ground microbial associations are linked and their relative importance for allowing native plants to adapt to rapidly changing environments are unknown. Here we examine interactions among an above-ground fungal endophyte that grows systemically in grass leaves and whole soil communities collected from the field. In the first of a two-part study, we examined three way interactions among the native grass Agrostis hyemalis, the above ground fungal endophyte Epichloë amarrillans, and a native soil microbial community along a water availability gradient. In a full factorial greenhouse experiment we planted naturally endophyte infected (E+) or experimentally disinfected (E-) plants in pots inoculated with “live” or sterilized soil, and subjected them to four watering levels. In part two, we examined how the endophyte and watering history influenced plant-soil feedbacks. Pots were inoculated with soil conditioned by E+ or E- plants from the highest and lowest watering levels in the first study, and seeded with 50:50 mixtures of E+ and E- A. hyemalis.Results/Conclusions  Endophyte infected plants produced 33% more flowering tillers at low water availability relative to uninfected plants. This benefit disappeared when water was not limiting. No differences between the live and sterilized soil treatments were observed for any plant responses. This is one of a few studies conducted to show endophyte benefits via drought tolerance in a native grass species. The effects of abiotic conditions were modulated by the above-ground endophyte but unaffected by the presence of below-ground microbes showing that the above-ground microbe trumped the below-ground microbes in altering plant response. In a surprising finding from the second study, soil conditioned by E- plants reduced the percentage of endophyte infected seedlings by 15%. There was no effect of the watering history on percentage of seedlings infected. These results suggest that soil feedbacks can alter the relative abundance of aboveground microbes through differential performance of the host. Climate models predict net losses in soil moisture for many regions of North America, yet little is known about the role microbial associations may play in the adaptation of plants to changes in the abiotic environment.