RO membrane colloidal fouling experiments were performed in the laboratory under well controlled and realistic conditions. Iron oxide was selected as a typical inorganic colloidal foulant, due to its importance, as evidenced from well known manufacturer recommendations on iron concentrations in feed waters and from frequently encountered problems in membrane installations. A range of iron concentrations was identified where a linear relationship existed between flux reduction rate and concentration. The performance of the Silt Density Index (SDI) was tested on the basis of the RO fouling data obtained. The range of iron concentrations where measurable and meaningful SDI values could be obtained was remarkably close to membrane manufacturer recommendations. A notable sensitivity of the SDI was also observed with particles for which retention is negligible. However, on the basis of the RO fouling data obtained, it appears that the SDI is not conservative enough. Furthermore, since the SDI cannot predict fouling rates, it cannot discriminate between different types of membranes.