The colorimetric Fe2+ indicators bathophenanthroline disulfonic acid (BPDS) and 3-(2-pyridyl)-5,6-bis(4-phenylsulfonic acid)-1,2,4-triazine (FZ) are routinely used to assay for plasma membrane ferric reductase activity in iron-limited algal cells and also in roots from iron-limited plants. Ferric reductase assays using these colorimetric indicators must take into account the fact that Fe3+ chelators (e.g. ethylenediaminetetraacetic acid) can also in general bind Fe2+ and may therefore compete with the colorimetric Fe2+ indicators, leading to the potential for underestimation of the ferric reduction rate. Conversely, the presence of BPDS or FZ may also facilitate the reduction of Fe3+ chelates, potentially leading to overestimation of ferric reduction rates. Last, both BPDS and FZ have non-negligible affinities for Fe3+ in addition to their well-known affinities for Fe2+; this leads to potential difficulties in ascertaining whether free and/or chelated Fe3+ are potential substrates for the ferric reductase. Similar issues arise when assaying for cupric reductase activity using the colorimetric Cu+ indicator bathocuproinedisulfonic acid (BCDS). In this paper, we describe an oxygen-electrode-based assay (conducted in darkness) for both ferric and cupric reductase activities that does not use colorimetric indicators. Using this assay system, we show that the plasma membrane metal reductase activity of iron-limited cells of the green alga Chlorella kessleri reduced complexed Fe3+ (i.e. Fe3+ chelates) but did not reduce free (non-chelated) Fe3+, and also reduced free Cu2+ to Cu+, but did not reduce Cu2+ that was part of Cu2+ chelates. We suggest that the potential for reduction of free Fe3+ cannot be adequately assayed using colorimetric assays. As well, the BPDS-based assay system consistently yielded similar estimates of ferric reductase activity compared with the O2-electrode-based assays at relatively low Fe3+ concentration, but higher estimates at higher Fe3+ concentrations with chelators other than desferrioxamine mesylate. With respect to cupric reductase activity, the O2 electrode consistently provided much higher estimates; we suggest that this was as a result of Cu2+ chelation by BCDS leading to a large underestimation of the true cupric reduction rate. These results suggest that an O2-electrode-based metal reductase assay system has some specific advantages compared with the traditional colorimetric assay system, including especially the ability to discriminate between the reduction of free metal ions and chelated metal ions.