SynopsisItapid membrane osmometry, narrow distribution polystyrenes in the range 4000-700,000, and three membrane types of different porosity were used to study the effects of diffusion and of Staverman membrane selection in osmometry. It was found that equilibrium was not obtained in the rapid measurement with the two densest membranes which led to grossly low values for the high molecular weight sample but had small effect at low molecular weights. Because of membrane selectivity effects at lower, and of nonequilibrium effects at higher molecular weights, the effective range was about 4000-350,000 for the densest membrane, 90,000-500,000 for the medium porosity membrane and from 146,000 to an unrestricted tipper limit for the grossest membrane. A unique characlerislic of the Mechrolab 601 instrument used in the study is the small constanL volume on the solvent side of the membrane which allows determination of t,he osmotic pressure of t,he diffused material and its concentration when the molecular weight is known. Iliffusion can be avoided if t,he pressure reading is taken quickly enough but membreiie selectivi1,y effeck cannot8 be avoided if diffusible solute is present. Enormous seleclivi1.y effect,s were found when mixtures of diffusible and nondiff wible molecules were measured. Experimentd value-; of the selectivity coelficient S, of the mixtures were Eoiuid t,o agree closely with those calctilaLed from the Staverman equat,ions. Estimates of the coil size of the polymers are given to provide an estimate of the effective porosities of the membranes.