Electron crystallography is becoming a powerful tool for the resolution of membrane protein structures. The past year has seen the production of a bacteriorhodopsin model at 3.5 A and the structure of aquaporin 1 approaching atomic resolution. Determination of surface topographies of 2D crystals using the atomic force microscope is similarly advancing to a level that reveals submolecular details. As the latter is operated in solution, membrane proteins can be observed at work.