Liposome structure and solute entrapment in multilayered vesicles (MLVs) prepared by reverse-phase evaporation (REV) were studied. MLV-REV vesicles prepared from ether/water emulsions have high entrapment. Entrapment depends on drug, drug concentration, lipid, lipid concentration, and the container used to prepare the vesicles. By use of 300 microL of aqueous phase and 100 mg of phosphatidylcholine (PC), vesicles prepared in a test tube 25 mm X 175 mm have higher entrapment than vesicles prepared in a 100-mL round-bottom or pear-shaped flask. By use of a test tube, 100 mg of PC, and 300 microL of aqueous phase containing sucrose (1-50 mg/mL), greater than 90% sucrose entrapment was obtained. Increasing lipid content to 150 mg of PC decreased entrapment to approximately 80%. Neutral PC MLV-REV vesicles have optimum entrapment. Mixing negatively charged lipids or cholesterol (CH) with PC to make MLV-REV vesicles results in decreased entrapment compared to using only PC. Preparing vesicles with the solid lipid dipalmitoylphosphatidylcholine (DPPC) or DPPC/CH mixtures (0 less than or equal to mol % CH less than or equal to 50) results in approximately 30-40% entrapment when diethyl ether is used to make the MLV-REV emulsion. Substituting diisopropyl ether for diethyl ether and heating the MLV-REV emulsion during vesicle formation generate DPPC/CH vesicles that entrap 60% of added solutes. The high entrapment found for MLV vesicles prepared from water/organic solvent emulsions depends on maintaining a core during the process of liposome formation. A method to calculate the fraction of water residing in the liposomes' core is presented and used to compare multilayered vesicles prepared by different processes. X-ray diffraction data demonstrate that a heterogeneous distribution of lipid may exist in multilayered vesicles prepared by the REV process.