A series of experiments were undertaken to investigate the effect of ionic strength and the concentration of free sodium ions in the resin gel on the performance of the diffusive gradients in thin films (DGT) technique. When the free sodium ion concentration in the resin gel was estimated by the time-dependent release into solution, it agreed with a previous estimate. However, equilibration with different volumes of water gave a higher value, suggesting that inherent averaging in the time-dependent release method underestimates the free concentration. DGT measurements of Cu and Cd were made over a wide range of ionic strengths (from 3 μmol l−1 to 0.8 mol l−1). For all the ionic strengths above 100 μmol l−1 there was no significant difference between measurements made by DGT and measurements made directly on the solution using atomic absorption spectroscopy. Below 100 μmol l−1 results were erratic. They did not comply with a theory that predicts high results for DGT based on enhancement of the diffusion coefficient of trace metal cations by counter diffusion of sodium ions. When Cd in solutions with a range of ionic strengths was measured by DGT there was no difference whether the resin gels were in Na or Ca form. Rather than counter diffusion of Na ions, it is suggested that the spurious behaviour at low ionic strength is due to interactions of the trace metals with the diffusion gel when there are insufficient excess cations present.