Concentrations of weak LY2109761 functional groups on HA were 5.5 × 10−3 M and 7.7 × 10−3 M for the G-HA and P-HA solutions respectively calculated from the total acidity as measured by Marshall et al. (1995). Concentrations of EDTA carboxyl groups in the extracting solutions ranged from 2 × 10−3 M (in 0.0005 M EDTA) to 0.2 M (in 0.05 M EDTA). Therefore, it is reasonable to expect that the highest concentrations of EDTA should have enabled complex formation with all labile Cu whereas at the two lowest EDTA concentrations HA should have been a strong competitor for Cu binding. Fig. 4 shows 63Cu and 65Cu chromatograms for P-HA in 0.0005 M and 0.05 M EDTA. It should be noted that the (native) 63Cu present originates solely from residual Cu in the ‘purified’ HA which has either resisted acid dialysis during the preparation of the HA or been adsorbed from distilled water during the extended dialysis against water used to remove residual acid (Marshall et al., 1995). The 65Cu was equilibrated in the HA solutions for 40 d prior to extraction for 2 h with EDTA. A decrease in ICP signal was observed for both isotopes bound to HA, with an increase in the EDTA-bound metal peak, as the EDTA concentration was increased. However, even though it might be expected that 0.05 M EDTA would be able to complex all Cu in the system, a significant proportion of Cu (>40%), for both residual (e.g. 63Cu) and spiked (65Cu) isotopes, remained bound to both HAs. This provides further evidence of non-labile Cu bound to humic acid. The diffuse peak representing Tris-Cu complexes (c. 15–19 min) also diminished, as the EDTA concentration increased. The EDTA-Cu peak shifted to a longer elution time at higher EDTA concentrations which may indicate changes in speciation during elution within the SEC column.