A key property of metallo-proteins and -enzymes is the affinity of metal ion M for protein ligand P as defined by the dissociation constant KD = [M][P]/[MP]. Its accurate determination is essential for a quantitative understanding of metal selection and speciation. However, the surfaces of proteins are defined by the sidechains of amino acids and so abound in good metal ligands (e.g., imidazole of histidine,thiol of cysteine, carboxylate of aspartic and glutamic acids, etc.). Consequently, adventitious binding of metal ions to protein surfaces is common with KD values > or = 10(-6) M. On the other hand, transport proteins responsible for 'chaperoning' essential metals to their cellular destinations appear to bind the metal ions selectively (KD < 10(-7) M, both for speciation and to minimise the toxic effects of 'free' metal ions. These ions are normally bound with still higher affinities at their ultimate destinations (the active sites of metallo-proteins and -enzymes). This review surveys possible approaches to estimation of these dissociation constants and pinpoints the various problems associated with each approach.