Abstract In this paper, representative polymeric (a PSf/PVP membrane), ceramic (a ZrO 2 membrane) and organo-mineral (a ZrO 2/PSf membrane) ultrafiltration membranes, all in the tubular configuration, are being compared for their basic membrane properties, and for the typical ultrafiltration application of protein recovery of cheese whey. These three different membranes with a quite similar pore size (the cut-off values for each of the three membranes were comprised between 25 000 and 50 000 Dalton) showed pure water permeability coefficients between 135 and 1250 l/h m 2 bar. The highest pure water flux was found for the organo-mineral membrane, the lowest for the polymeric membrane. By FESEM analysis of the top-surfaces (skin) of both the PSf/PVP and the ZrO 2/PSf membrane a strong difference in surface-porosity was found. These results were claimed to partially explain the difference in pure water flux. From SEM pictures of the cross-section of the ZrO 2/PSf membrane it could also be seen that the skin layer thickness is smaller, at these places where particles are present near the skin-surface, compared to the rest of the membrane as well as to the skin of the PSf/PVP membrane. These latter observations were also used to further explain the flux difference between the PSf/PVP and the ZrO 2/PSf membrane. In application tests (ultrafiltration of a sweet Gouda cheese whey) these three rather different membranes surprisingly showed practically the same gel-layer controlled or plateau fluxes, the same flux stability, and flux/concentration factor behaviour. The protein retention in all the experiments was 99% or more. The permeability coefficient however for this sweet Gouda whey was identical for the PSf/PVP and the ZrO 2 membrane and equal to 50 l/h m 2 bar. On the contrary for the ZrO 2/PSf organo-mineral membrane a nearly three fold higher permeability coefficient of 135 l/h m 2 bar was found. This property is partly attributed to the much higher surface porosity of the organo-mineral membrane as compared to the polymeric membrane. From this comparison one may conclude that for high fouling applications, the only positive effect upon using membranes with high permeability coefficients is a reduced transmembrane pressure for a given flux. However, for low fouling applications distinct gains in terms of flux can be expected upon using such membranes.