Arsenic compounds have been applied at the land surface as pesticides in agricultural areas globally. The purpose of this study was to evaluate the fate of anthropogenic arsenic applications related to agriculture, using arsenic applications on cotton in the southern High Plains (SHP), Texas, as a case study and examining possible linkages with contamination of the underlying Ogallala aquifer in this region, where 36% of wells exceed the new EPA 10 microg/L standard. Unsaturated zone soil samples were collected from boreholes beneath natural ecosystems (grassland/ shrubland) to provide a control (no arsenic application) (5 profiles) and cotton cropland (20 profiles) for analyses of water-extractable arsenic, vanadium, phosphate, chloride, and nitrate. Natural ecosystem profiles have high arsenic concentrations at depth (maximum of 7.2-69.6 microg As/ kg dry soil at 5.9-21.4 m depth) that are attributed to a geologic source. Most profiles beneath cotton cropland have high arsenic concentrations within the upper meter (profile means 1.7 to 31.6 microg/kg) that correlate with phosphate (r = 0.70, p < 0.01) and are attributed to anthropogenic arsenic application associated with phosphate fertilizer application. High arsenic concentrations at >1 m depth (profile means < or =36.3 microg/kg) found in cropland profiles are attributed to a geologic source because of similarity with profiles beneath natural ecosystems, lack of correlation with phosphate, and pore-water ages that predate anthropogenic arsenic application in many profiles. GIS analyses showed poor correlations between groundwater arsenic and percent cultivated land (r = -0.15, p < 0.01), groundwater nitrate (r = 0.30, p < 0.01), and water table depth (r= -0.31, p < 0.01), further supporting the idea that anthropogenic-derived arsenic in the shallow subsurface is not linked to groundwater arsenic contamination in this region.