Extrapolation from permeation data for pure solvents does not predict accurately their behavior in mixtures. In this study we examined whether the permeation of natural rubber, nitrile, and PVA glove materials by solvent mixtures is proportional to material swelling. Gloves were exposed to solvent mixtures and their permeation was monitored: breakthrough times, permeation rates, and degree of swelling were determined. Toluene exhibits higher permeation than MEK through natural rubber. Adding MEK to toluene decreased the permeation of toluene and increased that of MEK, proportionally to mixture composition and the degree of swelling. The opposite was observed for nitrile: toluene, now the "low permeation" solvent (LPS), retarded MEK permeation and vice versa. The effect was proportional to mixture composition but not consistently proportional to the degree of swelling. Samples were also exposed to the LPS for different time periods prior to a permeation run using the "high permeation" solvent (HPS). The decrease in breakthrough time of the HPS was proportional to the degree of swelling caused by the LPS. Material swelling appears to control MEK and toluene permeation through natural rubber. For nitrile, additional factors are apparently at work. No permeation was detected through PVA and material swelling was negligible. (c) 2005 Wiley Periodicals, Inc.