A batch, dead-end filtration method is used to compare the effect of various pretreatment operations on membrane performance. The dead-end method is compared with results obtained in cross-flow filtration as well as with conventional SDI measurements. Dead-end experiments were designed to approximate permeate flux over a range of recovery (up to 85%) while cross-flow experiments were operated at a constant recovery of 75%. Fouled membranes were imaged by SEM and, in some cases, EDAX to obtain information on foulant structure and composition. Among the advanced pretreatment processes evaluated, GAC pretreatment, combined with ozone, reduced fouling of both of the membranes tested to the greatest extent. UV irradiation and pre-filtration through a 5 mu m membrane did not measurably reduce fouling. The effect of GAC age on fouling was also investigated. GAC columns that had been in service for longer periods of time provided the same benefit in reducing fouling as did GAC columns that had been recently regenerated. This suggests that the primary benefit of GAC pretreatment may be related to biological activity on the GAC filters rather than adsorption of organic matter. Measurements of SDI, TOC values and flux decline suggested a relationship between feed SDI and TOC. However, flux decline was not predicted by SDI values.