Measurements of metals in soils by diffusive gradients in thin films (DGT) have previously been shown to be linearly related to metals measured in shoots of plants grown in pots. We examine the relationships between metals measured by DGT and other techniques with metals in the roots and unpolished grains of rice cultivated under field conditions at 18 sites in Jiangsu province, China. Rhizosphere soils of rice were collected and the concentrations of Cd, Cu, Pb, and Zn were determined on soil solution, acetic acid, and calcium chloride (CaCl2) extractions and by DGT. Simple linear regression analyses between concentrations of metals in plants and those measured using DGT and chemical extractions showed a very good fit for DGT measurements of the concentrations of all four metals in both rice roots and unpolished grains. Good fits were also found using soil solution and acetic acid extractions, but the correlation coefficients were lower than those obtained by DGT. CaCl2 extractions provided the poorest fits for all four metals. Multivariate analyses were used to assess the impact of pH, dissolved organic carbon (DOC), soil organic carbon (SOC), cation exchange capacity (CEC), and texture. Two principal components were extracted. The first was well correlated with SOC, DOC, and clay proportion and is therefore representative of "organic matter". The second primarily correlated positively with pH and negatively with CEC and is representative of "inorganic ions". When these principle components were included in multiple linear regression, correlation coefficients for plots involving metals in soil solution and in extractions using acetic acid and CaCl2 were improved, but there was little change in the correlation coefficients for comparable plots using metals measured by DGT. These results show for the first time that the DGT measurement quantitatively incorporates the main factors affecting bioavailability.