A device has been developed for the direct in situ measurement of metal remobilization into solution from settling particles in water columns, based on the technique of diffusion gradients in thin films (DGT). The device comprises a hollow cylinder with a DGT assembly at its base and acts as a traditional sediment trap, except that any remobilization from settling particles is reflected in the mass of metal taken up from solution by the DGT assembly. A control-DGT device (an upside-down trap) was used to measure the metal taken up by the assembly in the absence of remobilization from particles. Deployments of the control device under laboratory conditions and below the well-mixed surface layer of a lake showed that the cylindrical trap design, with a height : diameter ratio of 5 : 1, resulted in negligible turbulence near the base of the trap so that uptake by the control was governed by molecular diffusion. The DGT-trap and -control devices were deployed together within and immediately below the main oxycline of a seasonally anoxic lake in 1996 and 1997. The reproducibility of the metal uptakes measured by the devices in the field was typically better than ±10% for the controls and ±20% for the traps. The uptakes of Mn, Co, Al, Ba, and Ni, but not Fe and Cu, were significantly and consistently higher in the trap devices than in the controls. Reductive dissolution of Mn oxide appeared to be the source of the four remobilized trace metals, whereas Fe oxides made no significant contribution. The remobilization of Mn, Co, Al, Ba, and Ni was also reflected in elevated dissolved concentrations in the vicinity of the oxycline. Comparisons of the metal : Mn ratios obtained from the DGT-trap data with those calculated from the water-column concentrations indicated that the elevated concentrations of dissolved Co were derived solely from Mn oxide dissolution, whereas Ba, Al, and Ni also appear to be affected by other processes.