Virus removal using membrane filtration is designed to work by a size exclusion mechanism. Removal is dictated by the pore size distribution of the membrane and the virus size. However, information characterizing the pore size distribution of virus filters, and comparisons of quantitative predictions of virus retention performance from pore size distribution data to measured virus retention performance, is In this work, liquid-liquid porometry (LLP) was used to characterize the pore size distributions of developmental and commercial virus filtration membranes spanning a range of pore size distributions. Measurements were conducted using an automated high resolution LLP test system customized for virus filter membranes. Using pore size distributions determined from LLP, a mechanistic mathematical model grounded on the principle of particle retention by size exclusion closely predicted measured virus retention performance, both as a function of virus size and of membrane pore size. The data and model predictions from this study support the understanding that size exclusion is the primary mechanism for virus retention in membrane filters. An abridged LLP test, which is focused on a critical portion of the pore size distribution, was developed and is utilized at-line during membrane casting to assure consistent membrane pore size and therefore consistent virus retention. (C) 2014 Elsevier B.V. All rights reserved.