The properties of two supports, controlled-pore glass (CPG) and sand, and two methods, normal and ultrasonic, were compared for their effectiveness in the immobilization of the fungus Penicillium notatum and its use for preconcentration of copper, zinc, cadmium, lead and iron. The two supports and methods affected the performance of the immobilized fungus. With sand as support, the maximum preconcentration of the metal ions occurred at pH 9 for copper, zinc and cadmium, pH 6.5 for lead and pH 8 for iron using both methods, but with different Sensitivity according to the method used, With CPG support, similar behaviour occurred with both methods of preparation, and also with different sensitivities, and the pa;values at which maximum uptake occurred were 7.5 for copper and zinc, 8 for cadmium and iron and 6.0 for lead. Sand proved to be better than CPG for the preconcentration of iron and cadmium by the ultrasonic and normal method, respectively. With both methods and supports used, the immobilization procedure lowered the detection limits (from mu g/ml(-1) in the direct aspiration method to ng ml(-1) in the preconcentration method) and gave high enhancements factors (from 14 to 625 more than in the direct aspiration method depending on the metal). The highest enhancement factors were achieved for copper and zinc using fungus immobilized in sand support with the ultrasonic method. Further studies showed that interferences were negligible when interfering ions were in the concentration range 1-5 mu g ml(-1) for both supports and immobilization procedures. For higher concentrations (7.5 mu g ml(-1)) of interfering ions the main interference was due to Cd2+, especially in Zn2+ and Cu2+ determinations, when sand was used. Results of the determination of zinc and copper in tap and mineral water samples by this preconcentration method agreed with the results obtained by a reference electrochemical method.