In controlled metal sorption experiments, the equilibrium distribution coefficient is a key variable quantifying sorbate partitioning across the solid-solution interface. Separation of metals into 'dissolved' and 'particulate' fractions is commonly achieved with syringe filtration, where the boundary is somewhat arbitrarily dictated by the limited selection of available pore sizes. Investigations involving natural organic matter, such as bacterial cells or plant tissues, are especially prone to experimental artifacts if the substrate releases abundant colloidal compounds that contribute to sorption by binding free metal cations in a pH-dependent fashion yet pass through conventional filters, causing the truly dissolved fraction to be grossly overestimated. We observed this phenomenon during a study of lanthanide sorption on a marine macroalga, Ulva lactuca, as a function of pH. At low ionic strength, distribution coefficients calculated for a 0.22-mu m size cutoff falsely imply that metal sorption reverses to gradual release above pH 4.6, instead of continuing to increase. Centrifuging the filtrates in Amicon (R) Ultra units (30 and 3 kDa molecular weight cutoff) revealed a mounting proportion of colloid-bound metal, constituting up to 95% of the 'dissolved' (<0.22 mu m) fraction near pH 8. Measurements of DOC concentrations suggest this being due to pH-dependent binding of free metal cations to a fixed pool of organic colloids. The process is well described with a simple 2-site Langmuir isotherm in 0.05, 0.5, and 5.0 M NaCl. Using this model to correct the original distribution coefficients not only removed the spurious reversal at low ionic strength, but also uncovered a prominent suppressive effect on the intermediate and high ionic strength data that had initially gone undetected. Ultra-filtration may thus be an essential analytical tool for proper characterization and interpretation of metal sorption on organic matter over a wide range of experimental conditions. Some implications are discussed for the use of biosorbents in the remediation of metal-contaminated waste waters. (C) 2011 Elsevier Inc. All rights reserved.