Saccharomyces cerevisiae immobilized in agarose gel is proposed as a binding agent for the diffusive gradients in thin films (DGT) technique for determination of Pb in river water and seawater. DGT samplers were assembled with the proposed binding agent (25-mm disk containing 20%, m/v, S. cerevisiae and 3.0%, m/v, agarose) and a diffusive layer of cellulose (3MM Chr chromatography paper of 25-mm diameter). The effects of some DGT parameters (e.g., immersion time, ionic strength, and pH) were evaluated. Elution of Pb from the binding agent was effectively done with 1.75 mol L(-1) HNO(3). The deployment curve (between 2 and 24 h) was characterized by a significant uptake of Pb (346 ng Pb h(-1)) and good linear regression (R(2) = 0.9757). The experimental results are in excellent agreement with the predicted theoretical curve for mass uptake. Consistent results were found for solutions with ionic strengths of 0.005 mol L(-1) or greater and within a pH range of 4.5-8.5. Interferences from Cu (20:1), Mn (20:1), Fe (20:1), Zn (20:1), Ca (250:1), and Mg (250:1) in Pb retention were negligible. Determination of Pb in spiked river water samples (from the Corumbataí and Piracicaba rivers) performed using the proposed device was in agreement with total dissolved Pb, whereas measurements in seawater suggest that of the various species of Pb present in the samples, only cationic Pb species are adsorbed by the agarose-yeast gel disks. The in situ concentration of Pb obtained at two different sites of the Rio Claro stream (Corumbataí basin) were 1.13 ± 0.01 and 1.34 ± 0.04 μg L(-1). For 72-h deployments, a detection limit of 0.75 μg L(-1) was calculated. The combination of inductively coupled plasma optical emission spectroscopy and in situ deployments of DGT samplers during the 72-h period makes possible the determination of labile Pb in river water.