Soil contamination with arsenic (As), a carcinogen, is a global problem that stems from irrigation with arsenic-contaminated water, use of arsenical pesticides, fossil fuel combustion, and industrial processes including mining, smelting, and glass production. Plant-based removal of arsenic from soil, or phytoextraction, with the fern Pteris vittata L. is an emerging in situ technology potentially suitable for moderately contaminated soils. While theoretically promising, field application of phytoextraction with P. vittata faces limitations due to variable and slow arsenic uptake rates and long remediation times. Fertilization with nitrogen or phosphorus, and inoculation with mycorrhizal fungi, have been shown to increase arsenic uptake rates under simplified hydroponic or pot experiment conditions. It is unclear whether those findings might apply to field conditions where plant roots are not confined to a limited volume of growth media and where soil texture and arsenic concentrations could be heterogeneously distributed.In my dissertation I determined the effects of soil biogeochemical and physical characteristics, including 1) soil fertilization or inoculation with mycorrhizal fungus Funneliformis mosseae, 2) soil texture, and 3) soil arsenic concentration, on arsenic phytoextraction with P. vittata using a multi-scale approach. To test the feasibility of phytoextraction in fine-textured soils, I conducted a large-scale, 4-year field study at a prospective urban agriculture site (Berkeley, CA; 16-361 mg As/kg) working with community collaborators. To test whether leaching could limit the practical application of phytoextraction with P. vittata, I conducted a 1-year, high spatiotemporal resolution field study at a California superfund site (Richmond, CA; 25-114 mg As/kg). To determine mechanisms of arsenic mobilization and transport in soil during phytoextraction, I conducted a 0.4-year mesocosm study using novel unsaturated soil columns planted with P. vittata.Across all studies, I found that P. vittata was best-suited to phytoextract arsenic from nutrient-poor soils. Nutrient application had either no, or a negative, effect on arsenic uptake rates in P. vittata. In the 4-year field and mesocosm studies, uptake rates were negatively correlated with percentage clay, which was associated with increased nutrient content. In the 1-year field and mesocosm studies, I showed that arsenic was mobilized for uptake in P. vittata through ion exchange processes also known to mobilize phosphorus, and therefore, according to the phosphorus starvation framework, arsenic uptake could be a byproduct of nutrient scavenging processes most active in nutrient-poor soils. My results suggest that mycorrhizal fungi could increase arsenic tolerance through mechanisms other than nutrient transport. In the 1-year field study, inoculation with F. mosseae led to higher fern biomass, slightly higher uptake rates and significantly lower remediation times.P. vittata removed a mean of 3.6% of initial soil arsenic in the 1-year field study. Across the 1- and 4-year field studies, P. vittata arsenic removal rates ranged from 0.02±NA to 28.9±11.1 kg As/ha/yr, leading to highly variable and long remediation time estimates (12 to >1000 years). In the 1-year field study, soil arsenic concentrations were positively correlated with arsenic uptake rates but did not affect remediation times, indicating that a given soil was remediated equally efficiently regardless of arsenic concentration. Across the 1- and 4-year field studies, changes in soil arsenic concentrations were not consistent with fern arsenic accumulation and indicated complex cycling of arsenic during phytoextraction. In the 4-year field study, surface soil arsenic concentrations significantly increased in the coarse-textured soil, where greater fern growth and arsenic accumulation was observed, suggesting phytoenrichment of arsenic in surface soils over longer (>1 year) time periods. However, in the 4-year field study in the fine-textured soil, where fern growth was lower and mortality higher, arsenic concentrations were greater in the lower than surface depth interval, suggesting arsenic leaching. Furthermore, in the 1-year field study, soil arsenic concentrations significantly decreased, with a greater decrease in the lower (10-20 cm) than surface depth interval (0-10 cm). I quantified the first (to my knowledge) mass balance equating P. vittata arsenic uptake to significant decreases in soil arsenic, and found that arsenic uptake in P. vittata could explain only 38.5% of total arsenic depletion from soil in the 1-year field study. This discrepancy in the arsenic mass balance suggests arsenic leached in the 1-year field study, a finding I confirmed under mesocosm conditions. In the 1-year field and mesocosm studies, I showed that arsenic could leach as a result of rhizosphere processes mobilizing arsenic, processes which appeared primarily responsible for arsenic uptake in P. vittata. In both studies I found arsenic chemistry in the rhizosphere was distinct from in bulk soil, suggesting the rhizosphere was a unique reactive zone for arsenic mobilization. Furthermore, I found under mesocosm conditions that arsenic transported from bulk soil to the fern via transpiration flux of porewater accounted for