Sulfurospirillum barnesii is capable of anaerobic growth using ferric iron or arsenate as electron acceptors. Cell suspensions of S. barnesii were able to reduce arsenate to arsenite when the former oxyanion was dissolved in solution, or when it was adsorbed onto the surface of ferrihydrite, a common soil mineral, by a variety of mechanisms (e.g., coprecipitation, presorption). Reduction of Fe(III) in ferrihydrite to soluble Fe(II) also occurred, but dissolution of ferrihydrite was not required in order for adsorbed arsenate reduction to be achieved. This was illustrated by bacterial reduction of arsenate coprecipitated with aluminum hydroxide, a mineral that does not undergo reductive dissolution. The rate of arsenate reduction was influenced by the method in which arsenate became associated with the mineral phases and may have been strongly coupled with arsenate desorption rates. The extent of release of arsenite into solution was governed by adsorption of arsenite onto the ferrihydrite or alumina phases. The results of these experiments have interpretive significance to the mobilization of arsenic in large alluvial aquifers, such as those of the Ganges in India and Bangladesh, and in the hyporheic zones of contaminated streams.Sulfurospirillum barnesii is capable of anaerobic growth using ferric iron or arsenate as electron acceptors. Cell suspensions of S. barnesii were able to reduce arsenate to arsenite when the former oxyanion was dissolved in solution, or when it was adsorbed onto the surface of ferrihydrite a common soil mineral, by a variety of mechanisms (e.g., coprecipitation, presorption). Reduction of Fe(III) in ferrihydrite to soluble Fe(II) also occurred, but dissolution of ferrihydrite was not required in order for adsorbed arsenate reduction to be achieved. This was illustrated by bacterial reduction of arsenate coprecipitated with aluminum hydroxide, a mineral that does not undergo reductive dissolution. The rate of arsenate reduction was influenced by the method in which arsenate became associated with the mineral phases and may have been strongly coupled with arsenate desorption rates. The extent of release of arsenite into solution was governed by adsorption of arsenite onto the ferrihydrite or alumina phases. The results of these experiments have interpretive significance to the mobilization of arsenic in large alluvial aquifers, such as those of the Ganges in India and Bangladesh, and in the hyporheic zones of contaminated streams.