Coagulant-assisted direct filtration is a common process for treating upland, colored surface waters. This study investigated whether the process could be improved by partially replacing alum with cationic polyelectrolyte. Laboratory experiments were carried out to determine whether there is a unique optimal dosage combination of alum and polymer that--at a given color concentration and pH--maximizes the performance of a model filter. Filter performance was assessed in terms of the Ives filterability number and filtrate aluminum concentration. The results indicate that the filter performs systematically less well as the alum dosage is reduced. Partial replacement of alum with polymer appears to be less effective for coagulating color than alum alone at its optimal dosage. Includes 28 references, table, figures.