[en] Radiotracer experiments were carried out in 20 enclosures located in two lakes at the Experimental Lakes Area (ELA), northwestern Ontario, to study pathways of trace metal removal from the water column of shallow lakes. Two removal mechanisms were characterized: (1) sorption to and subsequent transport with falling particles and (2) direct adsorption to surface sediments. Our approach was to measure independently the kinetics of radiotracer sorption, fluxes and concentrations for particles, particle settling velocities, and the equivalent stagnant boundary film. Our radiotracer results enabled us to test the sensitivity of the tracer removal rates on these rate-determining processes using a numerical transport model. Acid titrations of whole ecosystems revealed that some trace metals (e.g. Mn, Co and Zn) can diffuse back to the water column as the pH is lowered from 6.5 to 4,8 after 18 d, while others remain tightly bound, (e.g. Sn, Fe, Se, Cr, Ag, and Hg isotopes). Subsequent CaCO₃ additions to bring back the pH to its original value restored the initial removal conditions for acid-sensitive radiotracers, indicating that the pH sensitivity is reversible. Transport parameters for particle-related pathways or diffusive pathways across the sediment-water interface obtained from our enclosure experiments were used to predict the removal rats of particle-reactive ⁶⁰Co and the diffusive pathway tracer ¹³⁴Cs observed in earlier experiments where radiotracers were added to whole lakes or to larger enclosures. (Author) (19 refs., 9 tabs., 11 figs.)