The technique of diffusion gradients in thin films (DGT) is designed to measure in situ labile trace metals; in a DGT device, the metal species are trapped in a resin gel comprising a Chelex resin in its Na-form after their diffusion across a diffusion layer comprising a polyacrylamide hydrogel. We deployed DGT and dialysis samplers (peepers) in the water column of the oligotrophic and acidic Lake Tantaré, Que. Canada. Contrary to expectation, concentrations of Cd and Ni measured by DGT in lake water were consistently higher than those measured by in situ dialysis. We show that co-diffusion of other ions through the diffusion layer, not taken into account explicitly in the original DGT theory, explains our field observations. When the DGT devices are deployed in a water of low major cation concentrations ( ∑[ cations]≤2×10 −4 M ), a steep negative concentration gradient of Na + across the diffusive layer as well as small positive concentration gradients of the other cations develop. The result is an increase in the effective diffusion coefficients of cations entering the DGT sampler and a decrease in the diffusion coefficient of Na + leaving the DGT device to fulfill electroneutrality. Non-steady-state conditions during the deployment time of the DGT devices make the prediction of these effects impracticable. We suggest that DGT devices, as presently designed, are not useful for measuring trace metal species in lakes of low cation concentrations. Such lakes are numerous in northern North America and in Scandinavia. Electrical effects on the diffusion coefficients were not observed in other studies using DGT samplers deployed in surface waters with major cation concentrations higher than 2×10 −4 M.