We compared the reactivity and microbial reduction potential of Fe(III) minerals in the rhizosphere and non-rhizosphere soil to test the hypothesis that rapid Fe(III) reduction rates in wetland soils are explained by rhizosphere processes. The rhizosphere was defined as the area immediately adjacent to a root encrusted with Fe(III)-oxides or Fe plaque, and non-rhizosphere soil was 0.5 cm from the root surface. The rhizosphere had a significantly higher percentage of poorly crystalline Fe (66??7%) than non-rhizosphere soil (23??7%); conversely, non-rhizosphere soil had a significantly higher proportion of crystalline Fe (50??7%) than the rhizosphere (18??7%, P<0.05 in all cases). The percentage of poorly crystalline Fe(III) was significantly correlated with the percentage of FeRB (r=0.76), reflecting the fact that poorly crystalline Fe(III) minerals are labile with respect to microbial reduction. Abiotic reductive dissolution consumed about 75% of the rhizosphere Fe(III)-oxide pool in 4 h compared to 23% of the soil Fe(III)-oxide pool. Similarly, microbial reduction consumed 75-80% of the rhizosphere pool in 10 days compared to 30-40% of the non-rhizosphere soil pool. Differences between the two pools persisted when samples were amended with an electron-shuttling compound (AQDS), an Fe(III)-reducing bacterium (Geobacter metallireducens), and organic carbon. Thus, Fe(III)-oxide mineralogy contributed strongly to differences in the Fe(III) reduction potential of the two pools. Higher amounts of poorly crystalline Fe(III) and possibly humic substances, and a higher Fe(III) reduction potential in the rhizosphere compared to the non-rhizosphere soil, suggested the rhizosphere is a site of unusually active microbial Fe cycling. The results were consistent with previous speculation that rapid Fe cycling in wetlands is due to the activity of wetland plant roots. ?? 2004 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved.