Recent advances in membrane introduction mass spectrometry (MIMS) are reviewed. On-line monitoring is treated by focusing on critical variables, including the nature and dimensions of the membrane, and the analyte vapor pressure, diffusivity, and solubility in the membrane barrier. Sample introduction by MIMS is applied in (i) on-line monitoring of chemical and biological reactors, (ii) analysis of volatile organic compounds in environmental matrices, including air, water and soil, and (iii) in more fundamental studies, such as measurements of thermochemical properties, reaction mechanisms, and kinetics. New semipermeable membranes are discussed, including those consisting of thin polymers, low vapor pressure liquids, and zeolites. These membranes have been used to monitor polar compounds, selectively differentiate compounds through affinity-binding, and provide isomer differentiation based on molecular size. Measurements at high spatial resolution, for example, using silicone-capped hypodermic needle inlets, are also covered, as is electrically driven sampling through microporous membranes. Other variations on the basic MIMS experiment include analyte preconcentration through cryotrapping (CT-MIMS) or trapping in the membrane (trap-and-release), as well as differential thermal release methods and reverse phase (i.e., organic solvent) MIMS. Method limitations center on semivolatile compounds and complex mixture analysis, and novel solutions are discussed. Semivolatile compounds have been monitored with thermally assisted desorption, ultrathin membranes and derivatization techniques. Taking advantage of the differences in time of membrane permeation, mixtures of structurally similar compounds have been differentiated by using sample modulation techniques and by temperature-programmed desorption from a membrane interface. Selective ionization techniques that increase instrument sensitivity towards polar compounds are also described, and comparisons are made with other direct sampling (nonchromatographic) methods that are useful in mixture analysis.