Cheatgrass (Bromus tectorum) has recently invaded marginal low-elevation salt desert habitats across the Great Basin. We tested the hypothesis that cheatgrass seed produced in populations from the more stressful salt desert vs. upland sagebrush habitats should grow differently in salt desert soils compared to adjacent upland sagebrush soil, and vice versa. We evaluated growth, incidence of flowering, and arbuscular mycorrhizal fungi (AMF) colonization of plants grown in the soils from which their seeds were collected vs. in the reciprocal soils from the nearest sagebrush or salt desert site in three large basins in northern Nevada. Simultaneously we measured nutrient cations, available nitrogen and phosphorus, percent carbon and nitrogen, texture, and dry-down characteristics in all soils. We found that salt desert soils were generally more nutrient poor and more saline than their upland (sagebrush) counterparts; salt desert soils also generally had a higher percentage of sand compared to their upland counterparts and were consistently drier. The most dramatic plant responses to soil and seed source were 1) lower aboveground biomass of mature plants in most salt desert soils compared to sagebrush soils, or lower biomass in plants grown from salt desert seed; 2) lower root:shoot ratios in plants grown in salt desert soil across two of three basins, irrespective of seed source; 3) a higher percentage of flowering individuals from salt desert seed sources at harvest, irrespective of soil source; 4) depressed AMF colonization of plants in salt desert soils; and 5) strong influence exerted by seed source on AMF, whereby sagebrush-originating plants grown in sagebrush soils had greater AMF colonization compared to salt desert soils but salt desert-originating seedlings had very low AMF colonization rates irrespective of soil source. These results suggest that both population level and soil-based controls are important as this widespread weed moves into marginal habitat.