Results are presented from laboratory batch tests using zero-valent iron to treat arsenic-contaminated groundwater. The laboratory tests were conducted using near-neutral pH groundwater from a contaminated aquifer located adjacent to a custom smelting facility. Experiments were designed to evaluate whether the permeable reactive barrier technology would provide a long-term, site cleanup strategy. The initial arsenic concentration was 48,000 ug/L; Zn concentrations were also elevated, up to 1000 ug/L. Arsenic was predominantly as As(III) with As(III)/As(V)=13.1 and no detectable concentrations of organic arsenic forms (i.e., DMA or MMA). Results indicate that zero-valent iron does attenuate arsenic concentrations from site groundwater to levels below 50 ug/L after 15 days (5 grams Fe/45 mL of solution). Both As(III) and As(V) were removed at comparable rates. During the batch runs, the formation of carbonate green rust occurred (XRD analysis) as the main surface precipitate. Extraction studies indicate that the green rust contains up to 1.6 wt% arsenic. In field applications of the zero-valent iron PRB technology, green rusts in addition to iron sulfides are common authigenic components. However, laboratory batch experiments are typically not run long enough for microbial communities to develop. Therefore, in a parallel series of experiments iron monosulfides were spiked into the batch reaction vessels. In these experiments, rates of arsenic attenuation are significantly greater than in experiments without the iron monosulfide spike (factor of 2.5). This result demonstrates the problems of extrapolating laboratory results to the field. Note: This is an abstract of a proposed presentation and does not necessarily reflect EPA policy.